Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
��������� � ���
���� ������������� !"#
��$% �� �&'()�*�+,-+�./!"#
��01 23
��������������� ���
�������� ���������� ���� ���
�� �������
���� �������
���������� �������������������� !�"
#$%&���'()*%+,-./%&0��1�23%&45��6789$:;
89�����<�=>*:�?@AB� ���C�DEFG ���H
I��JK�4LMNO�PQ%&R$�S�$�4T
UVW �� XY�(1) Z+[\2]�^_%&`abc�(2)����2]$[\
�Lde:��fg2]�(3) Ihij$kl89 mK 45 nopqr1�(4)
[\89<:;89�� qs�23�(5) ���� ��$89 45 st�u
AnLM�(6) ��vw xyz� FG�{+|}�+~Y���4T
�4�R�� #�����~Y JK��)�2]��D��6789<:;89
��������������|��-�&LMNO��������� H���
����)P��4T�� ����~Y���� ��&�������$�4D
E JK����0��1 ,�-������ �|+&T
�
���� ��������
� �¡����(��� "#$%&���¢£���� ¤F�¥¦%&�§¨©
(QªA)0��1 ����«¬���®���� �|�4T�4��§ ©¯�¢
£DE ����(�°±¤F�²)³��́ K� �&µ¶ ����$�|./%
&NO·���mK(Information Fusion) NO$�|�� ��¸5|+&T¹
�(��«¬<DARPA º5&R�� ����0��123�»¼��)�n½
¾���NO �¿$�)�ÀÁ� �+��(·���&ÂÃFGÄ$ÅÆÇÈ�
Generic �LMNO ��$��É|+�+T4$¯Ê����Ë=�ÌÍË$�|Î
~ �|+&0��1 �¤F��Google �CÏÐ Ñf?B@B��+|�Ò��
�����Ó¦��ÔÕÖ×-6789NO��+489�ÆØÙÚ�$&TR 45�
�������Ó¦%&ÔµÖ×���6789�:;89 ~Y�Ûeܯ�+T�
4�ÔÕÖ× 6789·��ÝW89 ÝW��Þß(��à��NO�Ü�-�Ò�
áÌBâK�Cãä�åæ�89 Ü'�|+&T:;�Ldeçè89�Z+[\
�Lde6789 é#��êë �|+& �����ìí�-åæ�NO��¡�
������%&R$�$C�É|+&T
R�� îï��
1. ðñ��~Y�ò5|ÏÐ Ñf?B@B�Ü�%&45������ ��$f
g2] Black Box$�)�:�89���*fg2]�óô-��+
2. µõFG Q/A$+Ø���� #ò ö÷P�Ì��ø¦ 45��FGùH
úû�:;�®�-�Ò�üý� ê+����É|+&
$+Ø���þ���� ö�$«¬ ðñ��t�B�B� ,���&· -�&T
¹ R����|��Ë=@A��-��UVW$+رñ��Ë=@A��ø ·$
��6789�:;89�st�uAnNO�?@AB�NO$+Ø���É4FG
���������� ����������� ��
���
�É)$�4��~Y�¿�é�&R$-��� ��NO�HI��JK�4
LMNO���%&R$�S�$�4T
R ��DEFG ��~Y������� ��<89<.� 45�JK�|�
�J��LMNO���%&����CREST ����Ý UV������-·Ûe
Æ'�|+�+· -�É4T
¹
���� �������
Ë=@A����9è$#�NO ���é��{e���VW$�����JK%
&R$-NO H�����%&��ÕVW������q�q �& Æ�Sv�4T
�ñ #�NO ����!i"`ab$��#^��$%&����89NO�!ii"
$¦FG %B�=@23$:;89�!iii"������� ��$89�&¯&
LMst�uAnNO�(iv)��vw ��'(�&¯&?@AB�NO� )¶
*+-��<���Æ�É4T
,-��#�NO ���~Y�/%T
�������� �� ������������������.67i��/0�`1��%
&P�2�o ¦3�~4��Þß�5%XY�%6|^_%&È_�7¶89-�
e�:á�(�;< �4P�2�o�=É|�{ { XY P�>·¥Ý�?
@%&R$ -�&T�4�:á�vw `1���.A��Generic �672�
o�B� :á��K'C|DEF%&NO�FGùH `1Þß�ij%&G1
�PQ�4TR�� 9�H�4`ab0��1 Enju ��ZI [\2]�
J%& Deep Parser -�(¶���������B���ÝW-ab-�&���ì
· -�&!�KLM�JK���Nâ"T`1 DEF��)�DO �-·�
BP���HE · $�É|+&!Q 1.3.1"T
�� ������� ��� �� ���� �����
�K� RB�
��>ÍB 76.9% 76.8%
ÏÐ 2�o 80.6% 81.0%
.S2�o(Feature forest) 88.0% 87.2%
�������
���������� �!�"# $$�%&���&�'(��� $)�%�*&+(��,-.-��/012345678�
9:;%�*&+(��-��<=>?��@ABCDE�FGHIJKHL7MNO�
�
� �!"#$%&'()*����+,-./01231�456���7��.T5U
�V Ì��·¶ XML �Ì���|·W��-Ñf�ÆØ*+XE0��1��
���!Y" 89XY�fg2]�óô C&R$-�v¦ [\2]Ñf�WZ
�Ñf%&0��1����4T�4�R��ÜÉ4JK���0��1$�|�
Y)[7 abÒ� Medline - Ñf0��1�23��BP�EBI(European
Bioinformtics Institute)-\� �|+&¥] 0��1�)^�4ÑfÈ_�7¶R
���������� ���������
���
$�P��4T
�8��9:;<=> �?� 2@1+A�B 456������.MEMM ��+��FG1
$�|�_Ùw<�`af G1�.S���� CRF(Conditional Random Field)$
¥b��+����F�Y�·�(¶��&(�WZ��H ÇÈ�R$��4�
c7de?@* NER �H�-�&R$����4T ���R ¥¸G1�c7
POS Ìf*ghii NER ����|�=jG1�)·^�4DO -&R$�
P��4TR�� ~Y��k� ��l$�mn �|�4Ihij$89no
pqr1 �o�Ì���pQ�µõ��þ��-JK%&· -�)�¬���·
W+qr�<|+&T
�C�DEFG4HIJKL1M�NOPQRS1���ATU L1M�����V���.Õs
sstu!Õv`�Usv7" AnnotatedText!GENIA :á�"�Î~����
���!w 1.3.1"�4TBP�Õ)s�x¯&��y1 R �Ì�ÜÉ|��
�|+&T�4�R �Ì��+e¶( ¬��0z�Ë- po½<�Ì
B{½$�|·®� �|+&TGENIA :á���[\Ì� ¶+4����
�� :á�-�&89-��e�` 2`2]<|e}W Nâ�o<ghB�
�(('&~��C�ãä��(��� ?@ �|+&R$(��»Z����¶
¶�&67�[\$:; ���{+|·�é#�L���$�&T
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�
�P ������� Q�RST UVW X/�YZ[\ �]^�
�
�W��X�QYZ4![\R1]1N9^)*4>_4`abcQdef ��g���h
��: WZ�=> ��-���$¶ uA���Ä n��������ÆØ
R$�e�1� JI- 600�@/� {uB=½ZO�EÝW�7%&R$�
~4�4T�4�� crawler� 20��O-����Æ��R �E�-�È ��
>�o�w�%&(ß 10,000�@/�)R$�P��4T�4��ä����-�&
�y�>������%&45 �o GPX ������� ����89�
��%&R$-� H���P��4!�KLM�JK���Nâ"T
cell line
artificial source
protein
peptide
amino acid monomer
DNA
RNA
polynucleotides
nucleotides
lipid
carbohydrate
other organic
compound
atom
inorganic compound
cell component
cell typetissue
organism
others
���������� ����������� ��
���
�i��j1klm4nop���q��.67�<����������&[� �� �
�$�|��������4�������W$= ��$ ÍBÌ>�0zB�¡
����4T$e��g�U�¢9$P�çè�ÜÉ4�W<�m£ÍBÌ>�0z
BÙÚ���������� = ����H%&R$-�� H���P��4T
g�U�¢9�ÜÉ4ÍBÌ>�0zBÙÚ��k¤ @A�y¥ [\d9�¦5
Æ'�+ãkl��@A�y¥��&ÍBÌ>�0zB������§9|�B�4
· -�&T
Ë=@A���� ÕVW��R�� #�NO89-�e�����JK%&��
·¿���çº��JK����*JK���0��1 23�ÆÉ4T,- JK
����̈ ¤ NO ¬�����©ª��&(�«¬%&· -�&R$����4T
�4�b���������4���0��1�23%&R$-���S� ®
!YKÕM"-¯64°Generic �6789�Lden½�¾�-�+LMNO-�(
¶���� ��#ò�R4¯&45 LMNO± ÇÈ�R$����4T
�r�stuv���w��.F²J³´ �o GPX�HPSG��&`abc(Enju)�I
hij��& POS/NER�µ¶·z¸:á�$+Ø�)¶ ��~Y�JK%&R
$��)�ghiiFG ������MedlinetuÛ�!Y)[7"��¹�W
-89%&R$�~4�4!Q 1.3.2"TR���Z+`ab�ÜÉ4��$�|��
=j ����<¡$·���e«¬%&· -�R ] 0��1-��Z Enju
�ÜÉ|·�F²89�Æ'�+ºK���ÕV,� ÝW�#%&T
�
�� ����-�� _`]abc<d?����ef�
[����]
PC��� �� 350 �����
IBM BladeCenter Cluster Dual Xeon 2.4GHz × 69 � Dual Xeon 2.8GHz × 42
Appro Blade Server Dual Xeon 2.4GHz × 64
���� �������� GXP (Grid Explorer)�
[����]
MEDLINE 1500��
!"� 7000� 14#$�
%&'(�)*(
�
�x�sty$z1{��������|��.÷P���#ò�7É4��!ghii
FG ��»"$¼½�¾¥���Æ�+��Ë=@A�� ~Y�(�� 45
���0��1$�|�B%&R$-��� H���/�4T¿�����ÀÁ
Â<Ã�Ä �o�À�ÅÆ%&0��1�ÇÈ�ÉÊÎ� #^0��1�Medline
:�Ñf0��1�Î~�4T�4����$¦��+ËÌ�7�;0��1(6
Í)*67��Ñf0��1(KIWI)�������4T
�
�
���������� ���������
���
� ���� ������
���� ����
���� ���� !"��
ÍBÌ���<ÎÏÐ�XQ �&JI���� ��Ñ¿<ÒÁ%&��
�����H��®�%&45���6789�%B�=@�Lde:;89�st
�uAnNO$+Ø�(�)�� ��&���HI���ÓF���J����<�
��ÆØ"# �&TµÀ�ÔÈeÀ¯&��·������o<n�Ë%&Õù�
NOLM �+$�Research Prototype$+Ø�Ö�×+4·��'&T
R R$(��UVW$+ØØÙ�±+��ñW��¯���o<n�Ë�Ú¯&
#�NO*qs� 23�Û��0|&�ñ�VW$��)JK�-��vw��
0��1���é��{e�ñÕVW�F9�5)�) �&�� Æ�Sv�4T
$e���ñ�����$¦¨� 45 n½¾���NO�����+45��!Y"
9è��ÜÛ�þ���LdeÝ� 6789NO�!Õ"�à��%B�=@89
45 ���qs� 23�!�"�¡�(¶�D��89�ÆØ45 JIÍ
Bt>NO�!)"��vw ��'(�Sv%?@AB�NO��Sv��)¶
�oË!6789µ�%B�=@µ�st�uAnµ�?@AB�µ" ��
Þ��ß0�4T�4�÷P���#ò�·¶����!ghii ��»"$¼½
���à_�ÆØR$-��ñ���B���ºÔ-���H��Ü�-�&JK0�
�1���%&R$� 94T
¹
�#��$%&'�
�}~��1���4����
¹ �� ��çº-��k¤ NOFG- �����ê+�á ·$�ÆØ45��
�ñ��âãäµå ��'Ò$æK'C ç�7��ñ����oËW ��
~Y����JK�4���ÆÉ4T� XY�st�uAnµ �=> ���
4�����K�ÜØR$-�6789 G *%B�=@G 67ij ���fg
2] 23�ÆÉ4T ���JK���{+|�ghii ������ Medline
!YUssvtu�Y)[7" %6|���|�[\2]�?@%&ab���Æ+�
� XY�������-�4 ��67µ�st�uAnµ�%B�=@µ ¼½
�à_ XY-�É4T
�����1�N4����
¹ ��� #ò�¨¯&45 LMNO23�Sv%�Ë=@A��-�����
$ ¾¥�� öÇî-�&T���è 0ì(��ghiiÄ ¨���$¶ é$
�|+4R$(��êë�<JBIRC�¬Q��i��ì�$¦��°íî1�Í?B�±
��»$¼½��á ·$����º54T$��JBIRC çº%& H-INV Ë=
@A��� Non-Funded Partner$�|Nï�4TJK���0��1 ÀÁÂ<Ã�
Ä �o�À0��1��(��$ ¾¥�� ~Y-�&T
���������� ����������� ��
���
�������
�Ë=@A��-��¬������q½%&R$�ð[��¬��0z�Ë
��ñ%&�C�'�'� ��~Y�¿���ò��4T�4�H_�óô�oË
(�B0o�õn���ÌBtö½��«¬<Parc�÷ByA�Ì��EBI�y·ø
ù�C)$¿�����à_�çº���� ��Ô- à_�o�PQ�4T4$¯
Ê�Coling(Õss))- GENIA��$%&�0z�Ë��úû����Ë=@
A��$�pQ-�& ��Ë=@A�� JSTü���ý ���þ��Y4�|{
)�GENIA:á� �������ª$�&�I�ÎÉ4T
�4��¬<ó��-�ñ�4°Z+67ab±�¶+| WS����B0o�õ
n�i<{�qB�i<�B�q�@�i�C�R FG H_���oË Nï�
<&R$-�R Ùw- �� ¬���®�F%&·$$�É4T� XY�¨¤
L�NO!��2]`1 P�2�o$����GENIA Ìf�c`abc" e
óô�oË ��-g� �����Ü'�|+&T
¹
�(��)*+,-�
¹ ¨¤ Sv�4Ùw!��>�o�6789���#^�{9&Z+[\89 ®
��ghii- ����÷ÍõB���C"��UV���G�%�&��# �
PF�|+�+�C�$�� �É4 �R UVW��� +Ò�· ���
Ò$�)���Þ���÷��4¬�ç��@¥�o$�l �� �&�Ø��
É|+&TLREC�Coling�ACL�HLT�SMBM �C�FG H_�¬�ç�- Lë
���y��qno����oË ��» �� �&�C�̈ ¤ ���oË
��R�� ��+ Ò�Â�%&Qº��É|+&T
4$¯Ê���>�o�6789��������� �¿$ GRID NO��)�
��UVW�»Z�� ÇÈ�$H�� �è �&�Ø��)�� 45 ���B
Ì!c¬.÷ByA�Ì�i�½Í�.Í?��i�«¬.:=>½�" øQ
�|+&��&+�� ��Ø$�|+&T
�4�ghii ����÷ÍõB�-��6789 ¬�ç�(Coling�ACL�C)
*ghii ¬�ç�(ISMB�PSB�ICSB �C)�$���0zB*�0z�Ë
�� �4T
��#^<��ÑfÄ Z+6789 ����"#�6789NO�·¶�oË
��+45�¨¤,ô-���B�q�@�<�o�i<?��B�>�<Parc<
�B0o�õn�<�ÌBtö½��C�Ó��|+&T48�R����J67
i�����q½�|�4�oË-�)����R ÙwÄ �� ïZ �&$
¦� �&T�+89$+Ø�n½¾���NO ÝX �����É)�4LM�
·¶NO ��ñ� É4TR Ò !F�·4��4µ¶ �'�'� �Ó�4�
�¯ �0z�Ë(2004V��¬<ó��)-�É4T
���������� ���������
���
����./&'����
¹
¹
¹¹
¹
¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹
¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹
¹
¹
¹
¹
¹
¹
¹
+,��%&-.$%&/0123/45�
6789:;<=>.$%&?@ABC�/45�
D�E�+�/45F �GH��I&�
JKL��MH?N@� �
�
� OPQ8�
� RS&T8UVWX�
� YZVW[�
Q!"\.$%&/]^/��_`abc�
�%&:defgL��MH?/VW:hi�
�
jk��G�N@� �
�
� OPQ8�
� RSlm�k��
� noVW[�
pqr$/stFjk��G�F/D��
pqr$:ru]vwxabcy�z�{|/
RS}~:hi�
VW����
��� ���
�
-�
-�
��GHk�N@� �
�
� P�Q8�
� RS8VWX�
� ��VW[�
�����/RSF��FRS��/VW�
.$%&N@� �
�
� OPQ8�
� RS&T8UVWX�
� ��VW[�
���������� ����������� ��
���
�� ������ ���� �
���� 01230453067+89:;<=>�?@AB��-C�
� � � � DEFGH� IJKLMNO�P�
���� �./QRST���
¹ WO���#^<��Ñf��B%&45 LMNO µ¶��å7�(�[\2]
�J%&`ab NO �&T$e�å�&2`2]89-�e��)Z+[\2]
�- J��J��ÆØ���67i��9è�´K�-�(¶�B� �����
B�&" #+67B���¨-�&WÚ��`12�o �� öÇî$�&TR
�Ø�9���`12�o ��-��(É445��mi��¨��é$%&��-
��$¦ FG$¨��$F�4�n½¾���NO Ü'�|�4$+¯&T¹
¹ �Ë=@A��-��!Y"B� :á��·$�`1�Î~%&R$�48��
Î~ �&`1��679è�´K��· -�)�(¶�� `1��$¦ FG�
� %&'(')-�& *+'+,(-�· -�&R$��4�!Õ"� `1��&`ab �B
� �¡���� 89�ܯ&W+89���·¶R$��Sv�|���ÆÉ|�
4T,--��`12�o 23$� WZ89 Õ®S�F9|��.%&T¹
¹
U�?@AB �8VWXN�8V-C 5Y�Z[�
¹ 679è�Ldeúû�2`2] ab���/67 [\�J%&45�"#ö
Çî�NO$�|�±ñW�'4)�� �|�4T679è-���W [\�`0
$�|Q^%&��b¤�2`�!,!1��23��oeM�C"�19|�� `
^B%&$�¯&T%�'�2`2] F(�Ê�C �Ø�2`�!, 4�4
( F(&45�` QB%&[\�J%&R$ ÇÈ$�&T
¹ 9è���$�|��Â5§6}2]`1(Head-Driven Phrase Structure Grammar,
HPSG)*77IÈ`1!Lexical Functional Grammar, LFG"�C�77F`19è$�Ê
�&679è #e�� ��µ¦ ~4��5|+&T���8+�R�� `19
è����:B9�ÌË=�>1$�|�H����� ���� 2`<[\2]
��6��J%&:�·;<�Æ'�|+&T�(��BP $R=��>*è`�
C �/67�����5��ab-�&2`ab0��1�?$<CÐP��+T�
�� �����ab%&45��� �|+&ÏÐ abc��67i��9èd9
�e�QI��@2]*åæ�o)192]�^_%&�$C�É|+&T
¹ ���-��úû�2 <̀[\2]�^_�(¶����2`abc���%&45�
Õ¶ §Ã��AS������aB%&G1�.S�4Tµ¶������`1��
�����%&45 Ù1è�·Øµ¶��D���Ì2]-QB �&2`2]�
��|P�2�o�ç¦%&45 nopqr1-�&T
¹ R�� G1��)�����c`ab0��1 Enju ���%&R$�~4�4
!w 3.1.1"TEnju� HPSG�Lde2`2]� ���ZI [\2]!¯7®2]"
��6��J%&2`abc-�&T�> ����(Wall Street Journal)��+4q
r�� XY�99.5% `���|2`<[\2]�^_%&R$ -���4�88%
���������� ���������
���
�O DO-å7W [\��o!¯7®�o"�^_%&R$����4TR �Ø
�úû�2`<[\2]�W+DO-J-�&2`abc����HE · $�É
|+&T
����> ����-�CõB��42`abc�ghiiè` �����
�¨ C&���ÆÉ4TEnju�Ý� c`abc-�& �DE�aF 45 P�
2�o� GENIA Treebank�ij�Ì$�|R�CõB� C4XY���+ij
�Ì��+4ºK-·WDO��~%&R$����4TR 2`abc��+|�
Medline Û�!YUssvtu�Y)[7" c`���� 2`ab�Æ+�� X
Y Ñf0��1G-�� �|+&Túû�2`<[\2]�^_��(¶R �Ø
�����Ì�89-�&2`abc����-�ì · -�&T
¹ ,--��`1��$P�2�oHIqr���¶+|�úû��.%&T
L1M�lm4 �56�
¹ HPSG��67 µõ��J���¦%&������$�å7ùH ��!de*
-KLMFtC1�C"�QB%&���$ ��K'C-��/67 2`<[
\2]�N÷%&`19è-�&T¡¯Ê�w 3.1.2 �O6e“choose” 77®S�/
%T“HEAD”�å7 de�Q��R ºK�6e� -verb$k¯ �|+&T“SUBJ”�
“COMPS”�R å7 Â7�S�7$�|$)Ø&å7 74�9�Ê���+2]
�Q%T“choose”�O6e� -���Þ�µ¶Ò¶|e!¶�)“HEAD” noun-�
&"�$��9�Ê���+$k¯ �|+&T�� 77®S���à��E�ÝÞ
*[\2]G��%&úû�Þß P�U���µ¶ 77®S�P�U��&Þß�
100,��HQR$·�&THPSG�Lde2`ab��`� å7 77®S 7¶
Þß�54%�Ø�2`@���Q|&R$-Æ'�&T¡¯Ê�w 3.1.3 �/%�Ø
��“they choose this strategy.”$+Ø`�ab%&$�“choose”�Â7$�|“they”�S
�7$�|“this strategy”�$&R$ -�&45�`Û� 2`2] R~%&Tw
3.1.3 �{+|�1 * 2 �!E�Q��¥¸ES 2]�¥¸Þß�74�9�Ê��
�+R$�Q�|+&T¡¯Ê�1 �TtheyU 77®S�k¯ �|+&Þß$
TchooseU Â7 Þß µV��9�Ê���+R$�Q�|+&T
�P ���������� �� �����
���������� ����������� ��
���
� HPSG���������� ���������������������
��� !�"#�����$%&'()*�+ �,-.-&��"/0�12
34567� �8��9:;&�<5=>���?@ABCD�EF�,G�H&I
J?��<�3K LMN1O�PQ�EF�RST���,HPSG �����<�"
/0�1234567�����?�9U�VW�$%�XY&��Z[�\]5W^
_`�ab��cdE�,��R�e�&VWf����5gh������<5i
j��R�G�ijk�W^_`flN�mlnR�op�qr5sT��<�tu
vw+&�&�,xyZ�?�E�VWz��5e��{|����<5ij��R�
G�ij�}~�G�8�67CD�EF��67� &�&�R]F��R��
�>���,W^_`flN�mln?@A������*��CDv�E~�G�
���$%&�5�����R� �*��E��G��+����*��CDv
5S�-���<5i�-+]��R?������?BCD�E�,
� �'(�?��<�3K LMN1O�PQ56��������<=>�e-]�
<�R-+���������(Miyao et al., 2003a, 2005; Nakanishi et al., 2004a,
2004b)5��-�,���<�?�W^_`5=>���*~�������5���
�,�N ���R?�"/0����-+�G�����67 5���sT�¡
!�E�,�&*¢���67�£6¤�¡ !�E~������"¤�¡ !�E
�,¥¦&9 §��N ���5=>���R� ���¨ 3.1.3 �©�}|�6
7 �ª«¬ W^_`�E�����N ���®�67 �ª«¬ 5¯�
��R�W^_`°±Z�����,
�l²³´�4�?���67�'(�µ�¶·§�+]� Penn Treebank5�HPSG
�N ����j¸���R�}~�9:;&¹W HPSG �<5=>-�º¨ 3.1.4»,
Penn Treebank?e¼ºWall Street Journal»�1234��-+����½Z& ���
¾¿5sT�� À3�E�,��5j¸���R�}~�HPSGÁ���� XY&
���� Ã8�� HPSG�N ���5���R��Ä-�,
8Å��<=>Æ? HPSG�Á��ÇF�lN�mln5ÈÉ��,Ê��ËÌ�Í
WNÎ 3ºPenn Treebank»5�HPSGÁ��ÇF��N ���ºHPSG�N ��
�»�j¸��,�<=>Æ�H&ÏÐ?��N ����j¸ÑÒ5��4² n�
��RR&�,G�Ó�HPSG�N ���®�ÔL fÕtuv?ÖlN�mln�
PHON “choose”
HEAD verb
SUBJ < >
COMPS < >
HEAD noun
HEAD noun
PHON “choose”
HEAD verb
SUBJ < >
COMPS < >
HEAD noun
HEAD noun
�� ��������������� ���PHON “they”
HEAD noun
SUBJ < >
COMPS < >
PHON “they”
HEAD noun
SUBJ < >
COMPS < >
PHON “this strategy”
HEAD noun
SUBJ < >
COMPS < >
PHON “this strategy”
HEAD noun
SUBJ < >
COMPS < >
1 2
PHON “choose”
HEAD verb
SUBJ < >
COMPS < >
HEAD noun
HEAD noun
1
2
PHON “choose”
HEAD verb
SUBJ < >
COMPS < >
HEAD noun
HEAD noun
1
2
PHON “choose this strategy”
HEAD verb
SUBJ < >
COMPS < >
HEAD noun1
PHON “choose this strategy”
HEAD verb
SUBJ < >
COMPS < >
HEAD noun1
PHON “they choose this strategy”
HEAD verb
SUBJ < >
COMPS < >
PHON “they choose this strategy”
HEAD verb
SUBJ < >
COMPS < >
�� ����������� ���������
���������� ���������
���
Wó±$+Ø,-�6��Ñ^%&R$ -�&T¶�)�`1��»�?>*ãµ
X� ¢Y��å�¥¦%&R$ -�����Z[�|+eR$-�)Wd���q
�B��Î)��|+eR$ -�&T%�'��Ù1è�{9&`1��$��
`19è�JÉ4ËqB0Ëo�54%�Ø��q�B��Î~�|+eR$�[
\%&T�G1��)�`1 µX� Þ\ ]^��)��4ÏÐ 67qs�
R��-�&45�`1�� :�� �"�_¯��&T
�#�0����-���
¹ ����2`ab��DE�aF "`-�&THPSG`1�`1��[�+2`2
]��^%&�Ø�ø �& �µõ���/67 `�`1��[�+2]�DE
7Ø&TnËq�0zB�Áä�DE� aF �4abXY�#ò%&T�4
É|�2`abc�Û| aab �(����abXY�^_%&R$ #ò �&T
�/6789-��P�2�o DE�aF 45�#e�+���~4��5|+
&T=É|�̈ ¤ §Ã���|·P�2�o ��ÇÈ$�¯��&T�(�� ��
HPSG`1�w 3.1.2�w 3.1.3�/�4�Ø�D���Ì2]!V?���2]"-
k¯ �|+&45�ÏÐ P�2�o���-��+TÏÐ �/6789NO��
`Û� 2]�¤F2]�Fa��� ¤F2]W J�pQ��c¦��¤F2]
P�2�o�ç¦%&�$+ØnË=y�$&T¡¯Ê�` de��å7l$
de�̀ Û� 2`@�µÖl$ Fd�Fa ��Fa �4¤F2] P��ç¦
��Û�2] P��¤F2] P� ¿$�|¦3%&TR�� G1����§Ã
2] @2]*àÄ2]-�&R$��.$�|+&THPSG�C-�+��&V?
���2]�e[ �>t2]-�&45�R G1���-��+Tµõ��V?�
��2] P�2�o�=jG1-ç¦%&$���-�+P�2�o <��&R$
:��|+&T
R §Ã��%&aBf$�|����-� feature forest �0� (Miyao and
Tsujii,2002)�.S�4TFeature forest$��vE�E�E @2]���42]-
QB%&µõ��Ì2]-�&T¡¯Ê�w 3.1.5�{+|�î½ di�DE��Q
��� Äî½ØC�(µ¶ Íg �&R$�Q%T=É|�w 3.1.5 feature
�P ����g������������ �! hi j�%Q klVmn[o�pq�
���������� ����������� ��
���
forest�ÕhÕhÕi¹k @2]�QB�|+&TFeature forest2�o��feature forest
��¦3 �4��?B�=92�o-�&T���-��feature forest���%
&R$�e��?B�=92�o á>�
Ì�ç¦%&6��nopqr1�.S
�4TR���)�feature forest î½ C
�Ø��Ì!V?���2]�C"-�É
|·�J�pQ� c¦����`�P�2
�o�ç¦%&R$ -�&T�4 É|�D
���Ì2]-QB �4~� P�2�
o�23%&R$ ÇÈ$�É4T
������-��HPSG 2`ab@*¯7®2] feature forest-QB-�&
R$�/�4(Miyao and Tsujii, 2003, 2005; Miyao et al., 2003b)TR����)�HPSG�
Lde2`ab 45 P�2�o�23%&Ù1�PQ�4T
�4�P�2�o�µ(�ij%& -��e�ÏÐ P�2�o�$¦FG ��
����¨ C&G1�.S�4(Hara et al., 2005)Tµõ��Ý� 2`abc�$¦
FG ����!ghiiè`�C"�� ����%&$�abDO j-%&R$
:��|+&T�(��kFGl$�����ij�Ì�Î~%& �B��-�
+T�R-�%q@�o P�2�o ���l�4���$¦FG�$F C&no
pqr1�.S�4T�G1��FG$���m�42`ab0��1�23%&
45 LMNO$�&T
���� `��� ���4u�N��uv�
���-.S�4G1��H��HPSG �Ldec72`abc Enju ���ÆÉ
4Tc7�>k~ �q�B�-�& Penn Treebank�����Ì$�|�2`a
b ���Æ+�Enju ���{�I.SG1 H���/�4TQ 3.1.1� Enju`1
��$2`ab�� XY�/%T
�Ò�Penn Treebankß)v`�HPSG�q�B��!(%&Ë=�>1��H��
�R(�ß 35,000å7���| 2,000ne 77®S�;<�4T�> ������
%&`1 op�!ab-�&` �K"�q¦�4XY�99.5% `���|r�(
3
d2
d1
d
1
c
2
c3
c5
c4
c7
c6
c
�P ����r��s� &����s(��t&�
å7E 34,754
77®SE 1,942
`op� 99.5%
¯7®�o abDO
!�K�ªRB�"88.0%/87.2%
Y`âãabÝW 0.56�
�� ����������� u�vwxyz���ef�
���������� ���������
���
�67×Ø5Ù����R� �8�G�|¢ 84.1%����-+?£6��� 5
����67��5����R5��-�,���?��8����67�?
��-�&]Û]vD5©-+]�,
Ê��feature forest ¡n5�·-�HPSG������67�ÜÝv6Þ ¡n5
=>-�,ÜÝv6Þ�×Ø�VW����ZopºbW_op»���5�È-�R
���88.0%/87.2%º��/ß�»���5��-�,XY&����5��}|&
Û]���Ù�� ���67Ú?/0�à����E�,8����67����]
+��$%&����5=>���R�}~����� 0.5�R]|67��5��
-+]�ºXY?Ê�5� »,
¦!�"�<��#v5á$�����GENIA Treebank% 4,500�5·]+ß&'-
� Enju�67��5� 3.1.2�©�,(N³)n� Enju�*++ 1,-�4.â��
�/â0������©-�e¼�1234�������1]��5��-+]�,
(N³)n� Enju�&'¡ !R*++ 1/10Ò��¡ !-.·]+]&]�(N
³)n�� ¡nR2Ȧ!�&'¡ !5�3�4����R�}~�5&]&'
¡ !��Û]��5�����R5"$-�,
�'(�?�"/0�12345 HPSG��� 67���R�6�+�Ä-�,"
·7�?�"/0�1234��-+XY&��z����°±Z��� �}|
�&F��R��$%&°ãÍW8Á9�"·:;§��,"Ó��l²³´�4�
?��<=&����"-� Enju 5·]+ Medline >?567��9:;&"@5)
F+]�ºXY? 3.2 �5� »,8��A'(R-+?��'(�}~�ÍW&��
��XY&��675"·BCn�m31DR�������Z&�<�5�E-�G
��#v5"$-�,
� �������
(Miyao et al., 2003a) Yusuke Miyao, Takashi Ninomiya, and Jun’ichi Tsujii. (2003). Lexicalized
grammar acquisition. In Proc. 10th
EACL. pp. 127-130.
(Miyao et al., 2005) Yusuke Miyao, Takashi Ninomiya, and Jun’ichi Tsujii. (2005). Corpus-oriented
grammar development for acquiring a Head-Driven Phrase Structure Grammar from the Penn
Treebank. In Natural Language Processing – IJCNLP 2004. LNAI 3248. pp. 684-693.
(Nakanishi et al., 2004a) Hiroko Nakanishi, Yusuke Miyao and Jun'ichi Tsujii. Using Inverse Lexical
Rules to Acquire a Wide-coverage Lexicalized Grammar. In the IJCNLP 2004 Workshop on
Beyond Shallow Analyses. 2004.
(Nakanishi et al., 2004b) Hiroko Nakanishi, Yusuke Miyao and Jun'ichi Tsujii. An Empirical
Investigation of the Effect of Lexical Rules on Parsing with a Treebank Grammar. In the
Proceedings of the third TLT2004. pp. 103--114. 2004.
(N³)n� Enju 85.40%/83.75%
GENIA Treebank ���&'
-� Enju
86.38%/79.96%
ß&'§�� Enju 86.82%/85.25%
�
�� ���������� !��"#$��%&�
���������� ����������� ��
���
(Miyao and Tsujii, 2002) Yusuke Miyao and Jun’ichi Tsujii. (2002). Maximum entropy estimation for
feature forests. In Proc. HLT 2002. pp. 292-297.
(Miyao and Tsujii, 2003) Yusuke Miyao and Jun’ichi Tsujii. (2003). A model of syntactic
disambiguation based on lexicalized grammars. In Proc. 7th
CoNLL. pp. 1-8.
(Miyao et al., 2003b) Yusuke Miyao, Takashi Ninomiya, and Jun’ichi Tsujii. (2003). Probabilistic
modeling of argument structures including non-local dependencies. In Proc. RANLP 2003. pp.
285-291.
(Miyao and Tsujii, 2005) Yusuke Miyao and Jun’ichi Tsujii. (2005). Probabilistic disambiguation
models for wide-coverage HPSG parsing. In Proc. ACL 2005. pp. 83-90.
(Hara et al., 2005) Tadayoshi Hara, Yusuke Miyao, and Jun'ichi Tsujii. Adapting a probabilistic
disambiguation model of an HPSG parser to a new domain. In Robert Dale, Kam-Fai Wong, Jian Su
and Oi Yee Kwong (Eds.), Natural Language Processing - IJCNLP 2005. LNCS3651. Springer-Verlag.
2005.
¹
������ ���� �������
¨¤�s-N÷�4Â5§6}2]`1 (HPSG) 2`abc (HPSG á�)
Enju WZF 45���1�y��&WZFG1$�Z+2`ab 45 b¤
�WZFG1 ���ÆÉ4 (Ninomiya et al. 2005)T�Ò CFG 45 �1�yn
opqr1$�1�y WZF$RB�w� 45 �C����2`ab
(Tsuruoka and Tsujii 2005a) � HPSG2`ab�����t+|����yA����
}2] _Þ�{�I CFGy¥B�á� (Tsuruoka and Tsujii 2005b) $ �Í�q
�½2`ab��&WZF��B�4T2`ab �Èqr��e�+��& Penn
Treebank (Wall Street Journal)������Æ+���Þ� G1�DO$abZO �
(�qr�4T
¹ åµF`1�67i�,ÚF$J�� u�(�#e���� �|�4T�(��
����ãä�DE�67i�2]�`���|@¯&· �� DE�k¯ 45
�J���ãä�v+$�� �|�4TWZF���nËq�0zB 45 Â
4&S� µ¶-�&45åµF`1 2`ab�� 45 �� ãä�e�
�|�4T��� ����)ab�� �"�wx �4· �2`abZO��
8����nËq�0zB 45��yF-��(É4T
¹ nV�#L�abÈ_�7¶åµF`1 45 P�2�o .S ��P�>�N
â%&R$��)��)����2`ab af�ÆØG1 ÇÈ� -|�4T¶�
)�µõ��·P�> W+2`@89 abXY$�|#ò �& -�����a
bXY�z{%&ÇÈ� W+¤Fab89J%�Ê�+TR Ù1��Ì�2`a
b*�1�y2`ab�XQ �&P�?� CFG 2`ab-#e�� �|�4T
¹ e åµF`1 45 P��2`ab �� �|�4 ���� e���
ÒÛ| 2`abXY�|}��ò5|{��� ��·É$·W+P� 2`@�µ
¶Íg%&�$+ØÙ1�$É|+4T�(��R�� G1 ���"/���ì
|}��2`ab ���mÐ�|��Ø$+Øðì �&TCFG ��� B��
�|P��?@��P�� CFG 2`ab��)�·P� W+2`@�Íg%&G1
·.S �|+& �R���P�� CFG 2�o���Ò�P��åµF`12�o
���������� ���������
���
-��+T
¹ ��-����:á�����åµF`1 P�2�o��+|�1�y2`
ab�ÆÉ4¡��e�̈ ¤����:á���%&P�� HPSG2`ab��1
�y���%&G1 �S�{�I� �È�¶+|�.�%&T�4�Z+2`a
b 45 ]¤ WZFG1 �È�¶+|·��-P�%&T
��. ��4>_4![���4���
�\��\�e�\�¹åµF`1-�& HPSG�Û| `1�J ���åµF Æ'
�&45�åµF WZF�2`abÛ� WZF���ez{%&T�Í��yA�
���åµF�~�%&��2] ��K'C�³�%&t�oÌqB�G1 µ¶-
�)�åµF�~�%&ÇÈ� W+á� (=��>�o�) �~��ò5|{��å
µF�ÆÝ������ á� > åµFÇÈ���ì�ë6&R$��)ö#�å
µF �Æ��9&R$ -�&T
� ¡./4¢£�¹�}2] _Þ��̀ �$`������+± 20å7,� }2
]�g~��+�$+Ø��q������-�&T¶�)�`� ���}2]�
����2`abXY�z{%&R$ �-�&$+Ø��J�¨��4G1-�&T
�¤�¥¦�§. ���¹�Í�q�½2`ab��)Z+2`ab�WZF%&G1
nV�. �|+&T�89$�|�RB��j+ �K� W+WZ��+2`a
b�Æ+�Z+2`ab afýW���e�)U£TabZO$abDO W���
C½%t �&45 Penn Treebank �Ø���� ����������|�Í�
q�½2`ab ���qr%&"# �&T̈ ¤���-N÷%&ãä�WZ-�K
� W+ CFGy¥B�á�� HPSG2`ab �89$�|�+4TCFGy¥B
�á���)^_ �4}2]�}´�$�|�+�� }´���4 &}2]
g~���%&R$��)afýW��"��)U£T
{1�z1��![��
�{2{f�¨¦©��¹ab�� }2]��µõ��<}2] � K3, }2]
� K3, }2]�Q%��2]>-QB �&THPSG 45 CYK2`abno
pqr1�{+|��Gr�}2] �K�µ¶ }2]���%&R$��)�J
¡ ����+-+&TR Gr�}2] �� ���}2]�K ·¶P�> Ø
�·W+P�>��� �&}2]�?@%&R$��)P�?� HPSG 2`ab
45 ��nopqr1 �B%&T��nopqr1-���a �� �|
+&T
�{1�z1��¹�1�y���a �� ��+(')��"�WZF��B%
&G1-�&T¥¸� K3�� K3 }2]�K �-�1) ·É$·P�>
W+}2] P�>(��1" w �g+4>,� P�>�7¶}2]�{�I
2) P�> W+ n k }2] ��ab ��$%&Tn�J��)ò54ô�P
���+|}2]Û� �-·É$·W+P�>(��1" w’ �g+4>,� P
���������� ����������� ��
���
�>�7¶}2] �ab ��$%& global thresholding·�H�qr�4T
���ª«�«¥. ���¹�1�ynopqr1-������) Æ'�a
g~ ��+R$ �&45RB� je�&T̈ ¤��abXY ^_ �&�-�
1"�#�� ��)��2`ab�ÆØG1�.S��H�4TCFG 2`ab
ºK��JXY�k��|{e:��$RJ:�� ?�Gr-�&45��)�
�J�ÆغK�,��J�4XY�Û|F�%&?Ø �+ �HPSG2`ab
ºK��åµF��&RJ�:�� W+45�,�J�4XY�R��%&T
���
���s-N÷ �4�Ø� Penn Treebank Section 02-22 ��+|g~ �4
HPSG`1��+�Penn Treebank Section 23,24�qr��$�4T2.4 GHz AMD
Opteron CPU���%&���Ü��4T�Ò��¯ Û| G1���K'C4
ëK�È�qr�4TQ 3.1.3� Section23 40å7,- `±��%&mÐ2] �
K��RB��F�:n(=�K�$RB� ë�âã)$µ`�4) âãabÝW�Q
�|+&T
¹ ���Ì���1�y�Í�C����2`ab �È�ØÙ�4TQ 3.1.4
��Section24 15å7,- `±��%&mÐ2] �K��RB��F�:n$µ
`�4) abÝW��Ì���1�y�Í�C����2`ab ØÙ�Q�|
{)�w 3.1.6�`±��%&abÝW ¡�E�>t-�&T
¹ �K�¹ RB�¹ ¢�:n¹ abÝW¹
Í�C����¹ £¤¥¦§¹ £¦¥¨§ £¤¥©§ ª¦« ¬q�`¹�
�� �������%�*&+(�-�"{g, |B)<=>?��x�}��
¹ �K�¹ RB�¹ ¢�:n¹ abÝW¹
�Ì�¹ ££¥®§¹ £¤¥¨§ ££¥¯§ ¯«ª°¨®ª ¬q�`¹
�1�y¹ £¨¥«§¹ £®¥©§ £±¥±§ ££ ¬q�`¹
Í�C����¹ £¤¥¦§¹ £¤¥®§ £¤¥©§ ¨¨ ¬q�`¹�
�� ����g��%�*&+(�-g"{�r |B)<=>?��x�}��
�
��
���
����
�����
������
�������
��������
���������
� � � � � �� �� �� ��
������ �����������
�������������
�
��
���
����
�����
������
�������
��������
���������
� � � � � �� �� �� ��
������ �����������
�������������
�P ����~��%�*&+(�-g"{�r |B)<=>?�
�������"�)x S&�� &+�� ����"�)�
���������� ���������
���
R Q(��̈ ¤ .S%& HPSG2`ab- �1�y ��G1����a�
^_%&�Ì�nopqr1$ØÙ�|�F�:n- 0.7%je�&89-�{�� 1000
² WZF��B�|+&R$ '(&T�4�¥b�¨¤ .S%&Í�C����
2`ab��)RB� {�� 5%?C�³��F�:n-{�� 2%?C�³%&R$
P�-�4T
¹ ����¯ ��Þ� G1 �Y�Section24 40å7,- `0���|qr�
4TQ 3.1.5$w 3.1.7�� XY�Q�|+&TQ�Z+2`ab 45 WZFG1
��+4ºK(Û| WZFG1)$�+�(É4ºK(�1�y �) XY�Q�
|{)�µ`�4){�� 2.5 ² WZF��~�|+&R$ '(&Tw ����
�1�y��&�Y�Q�|+|��1"�P�> "-ë´�4· (width)�
�1"�}2] E-ë´�4· (num)���� ��K'C(num+width)�global
thresholding (global)��C����2`ab(iterative)�global thresholding$�C�
���2`ab ��K'C(iterative+global)��+4��Þ� ºK abÝW$ F
�:n �o�Q�|+&TkG1���K'C�Ê��K'C&?C´+�Y�%�
' F�:n�µ$%R$�eabÝW ðe�É|+&R$ '(&Tw ���y
¥B�á�(chp)$��K'C4�Í�q�½2`ab �Y�Q�|{)�Û|
G1���|��Í�q�½F��&�Èw��P�-�4T
�¬6® ¯��
(Ninomiya et al. 2005) Ninomiya, Takashi, Yoshimasa Tsuruoka, Yusuke Miyao and Jun'ichi Tsujii.
(2005). Efficacy of Beam Thresholding, Unification Filtering and Hybrid Parsing in Probabilistic
HPSG Parsing. In the Proc. of IWPT 2005.
¹ �K� RB�¹ ¢�:n abÝW¹
Û| WZFG1¹ £±¥±§¹ £©¥®§¹ £©¥£§¹ ¹©«¤ ¬q�`¹
�1�y �¹ £±¥ª§¹ £©¥¤§¹ £±¥«§¹ ¯«ªª ¬q�`¹�
�� ����r����u��f�
���
���
���
���
���
���
���
���
���
�� � � �� �� ���
� �����������������������
�������
���
�����
�� !��
���"�����
���"�����"�� !��
������� �
������� �"�� !��
���
���
���
���
���
���
���
���
���
�� � � �� ��
� �����������������������
�������
���������
���������� ��
�����������!"#�!
�����������!"#�!� ��
������� �
������� �� ��
������� ���!"#�!
������� ���!"#�!� ��
�P �������%�*&+(�-g"{g, |B)<=>?���x � U��
���������� ����������� ��
���
(Tsuruoka and Tsujii 2005a) Tsuruoka, Yoshimasa and Jun'ichi Tsujii. (2005). Iterative CKY Parsing for
Probabilistic Context-Free Grammars. In Keh-Yih Su, Jun'ichi Tsujii, Jong-Hyeok Lee and Oi Yee
Kwong (Eds.), Natural Language Processing - IJCNLP 2004. LNAI 3248. pp. 52-60. Springer-Verlag.
(Tsuruoka and Tsujii 2005b) Tsuruoka, Yoshimasa and Jun'ichi Tsujii. (2005). Chunk Parsing Revisited.
In the Proceedings of the 9th International Workshop on Parsing Technologies (IWPT 2005).
��������������� !"��
¹ 67i9è�´K��`1�B���¨����%&�������èÝ UV��
�?$<C�e�leµ¤ ���oË!�B�q�@�i��B0o�õn�i�
?��B�>�i��o�i��¶·¸" ¨¤$¥Ýñ�����è�4$R=-
�É4T�� UVW A����~Y XY��Ø*e��)e ���oË
R Ùw- ���¹\�/�è5|+&T
¹ ��/�4¨¤ ��~Y��HPSG�L�$�4~Y-�& �LTAG �B0o
�õn�i�yyµ ?��B�>�i�º»µ Parc �{+|·�¥b�~Y�¼
�|�|+&T[\2]C�o- ab DO!¨¤ ºK���87%"��k�o
Ë$·¥�O$�É|+& ��4�FGÄ Tuning G1�WZ�nopqr1 �
�-��¨¤ �� µí�<¸|+&T$e��R Õ¶ NO �¨�Äw9|
½$�É|{)�ëK��JK��!�KUNâ"-·Y)[7 ���� R G1
-abÇÈ-�&R$ /C4R$(���)e ���oË ����¨��Z
+[\89�ÜØR$��&$¦��|+&T
¹ �Ë=@A��-��ghii è`*�>������F ÇÈ��/�4 �
Enju � Generic �`12�o$�É|{)� �¾��FG���ÇÈ-�&T���
¿çii��Ài�C�"#+FG ����89�ÆØLMNO$�&T
¹ 48�ÁÂ89*�o `89���%&���ã`1��` 89�C��)W
+Ú� "#$�&R$�R 45���[\<:;<`Ã�C��)¿���²)�
£�� "#$�&T67i��9è�´K�-�(¶�W+Ú��7¶ Enju ��R
�� ��)WO����ÆØ45 LM$·�&T
#$%� &'()�*+�),-./0123456-.�
� � � � 789:;� <=>?@�ABCDEFCGHIJKA@�ABL�
�M����NOPQCDEFC���
¹ ��������K(� :�Ñf���ÑfÝ�- 89$~�������KÛ
����|ÆÉ|{e89$�Ø�eFÄ%&"# �&T$e���½�
Ñf(�µíº<-���Å]��-Q É4Ñf�ÆØ��������~��8
9�����fg2]�óô%&R$-���%&�����K �U��ÆØR$�
�4�~� 89XY!�KY*�K� 89XY"�k��|{eR$-�ÑfÝ8
9�ÆÇ%&R$ öÇî$�&T~� 89XY��Èɺ�ÜÉ|2]F �&R
$(��XML Ì��?@<Ê�%&0��1�XML ������%&Ñf0��1�
{�I���2]�Ì����&:�����nËÍ� NO "#$�&T
���������� ���������
���
R�STU �VWAXYHZA��
:á� nî�0zB�ãä�����Ì_�#%&45�?��Ì*�
t=Í90��1�Cnî�0zB�]^�%&45 �o���%&R$�
ãä�é#-�&T$�nV�nî�0zB?� :á���+4�/6789
������ �4R$���%&$�� W+�G��&:á� ����·é#
�Ì��-�&$+¯&T�(��nî�0zB�o �� §Ã µ¶���
o� · �B��B�:�� W+�$+ØR$ �&TGUI*$Î�Ë=�>¬
B�67�p� API��+&$��o ��»,ô�� Ë=�>1 w´<�Ï
�%&R$ ãä�ÐÑ-�&T�4��B��B�:��89-��e�GUI��+
&Ë=�>1���:��·We�åæ��>u����%&89-·EÒä�#��
nî�0zB PÚ�!Ó��ÔÕ%&4I������:���"#$%&T
¨¤��XML -k¯ �4:á� �B��B��o����4T:á�
XML -k¯ �|+&$�%-�ÐP%& XML �Í9<89%&�o���-�
&R$��4���� �o�R���|���+�o�23%&R$ ÇÈ-�
)��k �B��B�:���aF%&89-��e���:�� Ô(�·ãä�
H�-�&T¡¯Ê�e uA��>u�� XML��Ö��|+&45�×�
�ÌÍo ø¦*!Ó��ÌÍo0�$�|.� ÇÈ-�)�R�� �>u��
µ(�Î)è5&"#��+T�4�O �oË��)�� �4 XML 45 Ø
�<tqst�·EÐP�|+&45R�� �o���%&R$·ÇÈ$�&T
w 3.2.1��������ÇÈ�Ì��nî���4· -�)�w 3.2.2����
ÙÚ%&45 CSS (Cascading Style Sheet)-�&TR��uA��>u��@¯&89
-w 3.2.3 �Ø�uA��>u�-ÙÚ%&R$ ÇÈ$�&
< ? x m l v e rs io n = "1 .0 "? >
< ? x m l-s ty le s h e e t ty p e = " te x t/c s s " h re f= "p ro te in .c s s "? >
< !D O C T Y P E s e t P U B L IC " - / /T M B O O K //D T D P ro te in A n n o ta tio n 0 .1 //E N " "p ro te in .d td ">
< s e t>
< s e n te n c e > A c tiv a tio n o f th e < p ro te in le x = "C D 2 8 _ s u rfa c e _ re c e p to r" s e m = "fa m ily " > < p ro te in
le x = "C D 2 8 " s e m = "m o le c u le " > C D 2 8 < /p ro te in > s u r fa c e r e c e p to r < /p ro te in > p r o v id e s a
m a jo r c o s t im u la to r y s ig n a l fo r T c e ll a c t iv a tio n r e s u lt in g in e n h a n c e d p r o d u c tio n o f
< p ro te in le x = " in te r le u k in -2 " s e m = "m o le c u le "> in te r le u k in -2 < /p ro te in > (< p ro te in le x = " IL -2 "
s e m = "m o le c u le "> I L -2 < /p ro te in > ) a n d c e ll p r o lif e r a tio n .< /s e n te n c e >
...
< /s e t>
��
�P ��-�����������V��n���
/* A S im p le C S S fo r P ro te in A n n o ta tio n * /
sen ten c e { fo n t-s iz e : 1 0 p t;}
p ro te in [sem = "m o lecu le"] { fo n t-w e igh t: b o ld ;}
p ro te in [sem = "fam ily"] { tex t-d e co ra tio n :u n d erlin e ; b ack g ro u n d -co lo r: c ya n ;}
p ro te in [sem = "co m p lex "] { tex t-d eco ra tio n :u n d erlin e ; b ack g ro u n d -co lo r: co ra l;}
��
�P ��-�-��������V��n�3��|�� �%%�
���������� ����������� ��
���
¨¤�R �Ø� XML 45 Ý��o���%&ÙÛ�$&TR�� ���
Eclipse20�-Æ'�4TEclipse�#e�i���I��êÜ�-�+��|+&%Ë
Bn���y¥�Ý��|{)�E Ë>�ÍB ÐP�|+&TRR-� VeX
XML editor21 :á�t�Ío�Ê�%&45�Ü'�4TDTD�@¯&R$��É
|�DTD-¦3 �4 XMLÌ� Íg*���?@ �4> Z[ ÇÈ$�&T
¨¤� XMLÌ���%&Ñf2@�o XConc�����Vex editor$ Xconc J
K?��Ì����4 (Kim and Tsujii 2005)Tw3.2.4�JK?��Ì Ü��Ô-�)�
TARGET=cons[@sem=”protein_molecule”], RIGHT=_word[text(“activate.*”)], UNIT=_word,
WINDOW=2-v¦ �&�?q ÑfXY�/�|+&TR �?q��)�cons
#� sem Þ� > “protein_molecule”-�)�� ���“activat.*” �&����
*+�Ñf%&R$ -�&TáÌB v¦��L���� XPath$¥¸0BÌ��
��Ý��|+&TXConc$ Vex editor �6%&R$��)ÑfXY��%&Z[�
C Ư&T
��
�P ��-��������\��/����
��
�P ��-�g���*$+�t� �� ����� +&(��¡�¢�(�*�
���������� ���������
���
¹ :á�nî�0zB�±ñW�'4&DE nî�Ì��&¾¥ÎÜ-�
&45�!Ó ßà$¾¥ÎÜ�Í9%&×�@zBÍ90��1 "#$�&T
CVS (Concurrent Versions System) �� ��S����40��1-�)�Eclipse�
CVS IÈ�RÛ��Ö��|+&45�JK?��Ì�·� IÈ ¥Ý�¶+
|e&TR �Ø�Ý� XMLtö÷����+&R$-�?��Ì�Ñf0��
1�×�@zBÍ90��1�Cá����:���#%&0��1�ðñW(¶j
+:��-��%&R$ ÇÈ-�&R$�/�4T
�
��������
(Kim and Tsujii 2005) Kim, Jin-Dong and Jun’ichi Tsujii. (2005) Corpora and their Annotation, in Text
Mining for Biology and Biomedicine, to appear.
�
��0123[E!()X\W�
`PÑf��e�+��&�qnBÑf-��å7 ^B�o� AND�OR�NOT�
���o-v¦%&R$ -�& �XML �XQ �&�2]F`P���|�2]
�o�v¦%&R$-��+T*+XE���qnBÑf IÈ�ï¯��2]F
������%&2] v¦�ÇÈ$%&Ñf 45 XE-�&T
ÏÐ *+XE-�2]W �o�v¦�4�?q�0��1�@¯4ºK��?
q$RÛ�÷�y%&`P*+ ��%45�?$<Ca <���+(�·�e�
âã?9 ��+�¡ a <��&( C�(��&R$ e�äå�Ñf�Æ
ØR$ Ñ�+T¨¤���?q�¤F�?q�Fa��k¤F�?q���
TFIDF�:n�@¯&R$��)�>B�?� XML��B�4 (Masuda et al. 2003a,
Masuda 2003b)T¤F�?q�Fa �|+&45��?q�RÛ�÷�y��+
ºK-·a <��&T�4�¤F�?q���| TFIDF ��&é�?9 � �
|+&45�XML ¤F2]���|·é�?9 �|+&R$��&T����
)��½ � Ñf*RÛ÷�y��&*+XE Ñf�)�>B�?� ¤F
÷�y��&*+XE ?Ø �)´+Ñf -�&R$�P��4T
�Ë=@A��-��s-N÷ �4 HPSG2`abc Enju ��)ab �4J7
2]{�ImÐ2]�Ñf��$�|+&Tw 3.2.5 �� XML ,Ú ^_ ¡-�
&TXML ��)J72] QB �|+& �mÐ2]$J72] Ì���Ì��
æ��4*+ 'ç%&R$ �&45�¥µ XML-QB%&R$ -��+T
w�/ �|+&�Ø��RR-�mÐ2]� ID��)QB �|+&TR �Ø
��Ì���|�ÏÐ *+XE��+|J72]$mÐ2]�v¦%&Ñf�ÆØ
R$�-��+T¨¤�*+XE�,- �Ø�ÔÕ��w-/ �& XML`P��
��Ñf -�&�Ø�*+XE{�I����%&nopqr1 w´�ÆÉ4T
1. ÏÐ *+XE��%&Ñfnopqr1�w´��ÏÐ nopqr1-�[�
eÑf-��(É4¥µÌ���&����B� ÐP%&ºK(¡.phrase Ì�-
æ��4*+ �� �� phrase Ì�-æ��4*+ ÐP%&)�{+|·WZ�
Ñf�Ư&�Ø��4T
���������� ����������� ��
���
2. *+XE !E��)�°¥µ 2]-�&±R$�/�|+& XM���|���
��v¦�|WZ�Ñf�ÆØnopqr1����4TR���)�°R 6e
Â7 R tCr-�&±$+É4�����-�Ä�|+&kì��&�o�W
Z�Ñf-�&�Ø��4T
D$-�������������K���bg ¡:�¢£F�b¤�¥¦§¨¥©
ª« activate /¬¥®$« proteinA¥ ¯°$« proteinB±²g ³:´f¥Fu¢ ¡
0µ:¶ 3.2.6.n/��B��b·�¸fg£F«±¹gº¶n±»¥(1)±¬¼« activate
/½$¥(2)¥(3)»¾¿À¿ proteinA¥proteinB :Á> phrase :�fºÂ|�Ã�$1¥$2�
b· proteinA:Á>�Ä�C« activate/®$(arg1)¥proteinB:Á>�Ä�C« activate/
¯°$(arg2)±²g£F:���ugº��BÅƱ»¾¿|ÇÈ/�K:fÉ�Á> :
���ugº
�����
[sentence] > ([word base="activate" arg1="$1" arg2="$2"] (1)
& ([phrase id="$1"] > proteinA) (2)
& ([phrase id="$2"] > proteinB) (3)
������������
A > B B��� A�������
A & B
A���� B������� !"�#�������� $�%!&�
'� and �(��
$1,$2 )*
[tagname]
tagname ��+�,�-.������/0�12�034�56!"�
78�9.�����
�P ��-�~��p£¤¥¦§¨£�[©lV���
¹�
�P ��-�r��ª�ef«¤¬®3 ¯� $+�� �� °\[°��
���������� ���������
���
¶ 3.2.7��¿g ¡��«ÊË/��B�����Ì|¿gº¶n/Ík�Î/�K»
(2)�Ï�Ð�]ÑÒ±²·¥ÓÎ/�K»(3)�Ï�Ð�]ÑÒ:��¥ÔÎ/�K»
“activate”$÷�y%&èì�Q�|+&T
��������
(Masuda et al. 2003a) Masuda, Katsuya, Takashi Ninomiya, Yusuke Miyao, Tomoko Ohta and Jun'ichi
Tsujii. (2003). A Robust Retrieval Engine for Proximal and Structural Search. In the Proceedings of
HLT-NAACL 2003 Short papers. pp. 58--60.�
(Masuda et al. 2003b) Masuda, Katsuya. (2003). A Ranking Model of Proximal and Structural Text
Retrieval Based on Region Algebra. In the Proceedings of the ACL 2003 Student Research Workshop.
pp. 50--57.
]�56-.^A_`A\4Wa\I�Abcd�
w 3.2.8 �6789�oË 0��1Û�é�Q�|+&TR �-2`ab 4
5 ���þ��Y4%HPSGá�Enju�2¶ L�0��1(�~É|{)�
µ¶� JSPSêj�ëË=@A��(q{: ìíîµ�1996-2000VO)-�� �4
è9VË=�>¬B�67 LiLFeS$� >BÌÍ1?B@B�·Øµ¶��Ë=@A
��-�� �4��2]���0��1-�&T
¨¤ Ý��4 HPSG���2]��%&Þß-¦3 �|{)��� `1���
��2]��%&åµFïÎ��)� �&TLiLFeS �V?��2]�L��Ì$
%&è9VË=�>¬B�67-�)�HPSG�k¯����2]�ïÎ%&2`ab
c�`1ijc�C 0��1 23�]^�%&T�(��è9VË=�>¬B�6
7 LiLFeS��Ë=�>1-k¯ �4�Ì*�2q� µÝ��Ì�ïÎ%&R
$�]^�-�& �ðt��t�Ío0��1��Ñ¿ �&�Ì���Í9%
&IÈ��B �|{�Ò��6��ij �45P®S�àñ�4)���2]-Q
B �&2`abXY�àñ%&R$�ÐÑ-�É4T
��
�P ��-����[©l±²³´ *&+� &�µ<=>?ef�
���������� ����������� ��
���
¹ ¨¤���Ò�C++(�ïÎ-�& Persistent Standard Template Library (PSTL)���
������+|�LiLFeS $ C++-ïÎÇÈ���2]�Ì��Í90��1�
���4T¨¤�����0��1 ���Sv�|{)�Ë=�ÌÍË0��1�
LiLFeS -k¯ �&R$ + �����0��1�Â� C++-R�H ��`1
���C ��2]ïÎ "#�ºK�� C++(� LiLFeS>BÌÍ12@�o ò
I^ �&T� 45 C++$ LiLFeS ó67(�]^�ïÎ-�&�Ì��Í9
0��1 ���+T¨¤��Ò C++-#eÜ'�|+& Standard Template Library
(STL) t�Ío0��1��Î~�4TSTL�q��*��Ì���0���q
�C µõ��:B��ÌÍË �Ì2]� C++�-]^�²)ôØR$�ÇÈ$
%&�BËC�>Í�>q-�)�C++ �ª>Í�>q$�|�e C++:Bá
Í>��[ �|{)��4�ãä�#eÜ'�|+&T�(�� ��t�Ío0�
�1��23 �& STL -�à��Ü��Ú¯Ø&��>�o��H�?$<C�
e�¨¤�t�Ío��ðt���Ì�àñ-�& STL (PSTL) �ø����4T
LiLFeS� �Ì��Í90��1� PSTL��+|�H �|{)���2]�à
ñ%&��Ì�q���{�I�Gr�V?��2]��$�|V?��2]�Ñ
f-�&��0�$�q �H �|+&T�4�åµFÇÈ�V?��2]��
$�|Ñf%&fgIÈ (Ninomiya et al. 2002)·��������H �|+&T
¹ fgIÈ��)���2]�Ì��(��åµFÇÈ�V?��2]�WZ�Ñ
f%&R$ ÇÈ-�)�����)�åæ�Ñf�ØÙ�Ey² ZO-Ñf-�&
R$ P� �4T��2]�>�o?��?�>t2]-�)��(��&�
- ��?@ �4>�o ��á�$ò<-+&TfgIÈ��+4Ñf��°åµ
FÇÈ���2]�Û| á����|{Ê+åµFÇÈ�> ?@ �|+�9�
MEDLINE�������
1200�� (15GB��
�� ��)
GENIA
����GENIA
������
��
��
���
���
HPSG����Enju
����
����
���
����
����DBMS
������������LiLFeS
HPSG
������
��
��
����
�� �
���
��
��
�
���
����
(Medusa)
���
����
�������
<E1���, �����>
<BBc���, ��>
…
…
…
…
“ ����”���
�“���”���
���
Since the PHO2
Asp-230 mutant
mimics
Ser-230-phosphorylate
d PHO2, we postulate
that only
phosphorylated PHO2
protein could activate
the transcription of
PHO5 gene.
�P ��-�¶��·Bbc¸6V¹�� ��º»¼�
�����
���������� ���������
���
Ê���+±$+ØåµF 45 "#���Ld+|+&T��2] �?q$�
|@¯��4�$��Ò�·É$·a ab���e�&R$ -�&á��6��E
kB¦�� ���� á� >(����a ab���e�)U£Ta�åµF�
�)ò5&R$��)�¤F�:;�k¯�4�?q�k¯-��Ñfâã�mÐ�
�+çè ÇÈ$�)�:�Ñf0��1$�| ´+����¯|+&$+¯&T¹
¹ HPSG 2`abXY���2]-QB �&J7<mÐ2]-�&45����
abXY�����2]�Ì���àñ��Ñf%&R$ -�&T̈ ¤���2
]��������?@%&0��1��S����4 (Ninomiya et al. 2005)Tw 3.2.9
�� ��w�Q�|+&TJ7ab�2`@�Q�|{)� ¤F��å7 *+
�¨�|+&TmÐ2]�J7ab }2]{�Iå7��¨�|{)�}2]�å
7W �o�/�|+&T�4�deÌf*ùH|abÌf XY·¥b����
����|?@%&R$ -�&45��¯ *+XE IÈ$��K'C&R$��
)�b¤�abc XY����4Ñf ÇÈ$�&TJK0��1-��R 0��
1��+|�Medline n���>��Û�(ßYUssvtu�ßY)[7)�ab�4
XY���2]�Ì���àñ�|{)�v¦ �4[\2]����1�,Å�
ÑfXY <��&R$�P��4T
��������
(Ninomiya et al. 2002) Ninomiya, Takashi, Takaki Makino and Jun'ichi Tsujii. (2002). An Indexing
Scheme for Typed Feature Structures. In the Proceedings of the 19th International Conference on
Computational Linguistics (COLING 2002). pp. 1248-1252.
nounsyntax
word ���
���������� �������������������� ����
�� !"#$%&'(�)*+,�-./�012��3�4
nounsyntax
word �����
verbsyntax
word ����
dtrs
npsyntax
dtrs
ssyntax
����
object
noneagent
predicate ���
modify
predicate ���
������
[ ]�����predicate
��
noun noun verb
�� ��
�5
�� ��������� �������������
���������� ����������� ��
���
(Ninomiya et al. 2005) Ninomiya, Takashi, Yusuke Miyao and Jun'ichi Tsujii. (2005). A Persistent
Feature-Object Database for Intelligent Text Archive Systems. In Keh-Yih Su, Jun'ichi Tsujii,
Jong-Hyeok Lee and Oi Yee Kwong (Eds.), Natural Language Processing - IJCNLP 2004. LNAI3248. pp.
197--205. Springer-Verlag.
���������� ������
� ��������� ���������������� ���� !"
�#$%&�'(� )'�*(Unstructured Information Management Architecture) +
,-.�/012�345� �6789 ::; <=&���>?@ <=�AB
CD�EF+G HIJKL6:M;�HIN���OP�QRSTUV#M�Q
W ;X6789 YZ�Q���[&�'(� )'�*\�Semantic Retreival System
M� ���[]�^� �6 _�`�a��Mbc� ^P�d>\ef� �6
ghiVPWjk-lm3n� �67
� : op <=Zq&��r�W�stuv�� YZwoxyz{�!"gh��
�[wox|z{ Q}�~� ��;��OW��� ���b�M�Q�";
��F\��5� �67����O ��; �� $%�#��-������
M-6:M\��;�67M���ABCD�EF+G����ghi&�)���(�
)Mu¡¢���£w¤ {M¥�;�O�> ����¦�����[wCheshire)
M Fm§¨�#:M©ª� �67Cheshire&�� �¤ �«¬��®¯
u�<=;W°��±�^P�°�<=²³�´µTU6:M\��;�67
���� ������������ !"#$%&'()*�
� � � � +,-.�� /0��1&234�
�5����6789"���
� /0�l¶£·&�����¸H¹º;-�����»®¼�½�¾¿�À�ÁÂ�
�<=-.�ÃÄ�Å °����ÆÇ5����-ZqÈÉ �67Q¹�st¸
H�����¸H�ÆÇL6Q}�&�:�> �� ÊË�X�QÌÍefL6Î
Ï\X67M���wÐ{�� Ã+GÑ��� Ò�\Ó�Ô-ÌÍw£·�ÏL6
ÕÖ�^V×�¸HØÙ{�wÚ{Ût n-ÜÝ V#-�� EF;�Op��Þ��-
����b� Robust-ÊFßLÌÍ�wo{st Êà;X6?áâ iã���
�ä�QÌÍ�wy{ståæ ��çuè[Ml¶£·ÌÍM�lN�Ý%"±é
6ÌÍ� ef\êëìM-67
� í<=¢Á¬îï�;&�ðñwÚ{ Q}�Ãòóôõ�}6Ó�Ô-�ö
(Maximum Entropy)÷�� Kazama 2005)�^V×�wy{M� �Enju �Ô÷��e
f�QwoxÐzøù{7: �ö÷��Ç� �Op�CRFM�ú-�BN-ûÆ
ÑüZL6Ó�Ô-ÌÍ(Easiest-First)ef�Q7
íz;&�ðñwo{�wy{ Q}�ef�QÌÍ�ý��þL67
:�;<=1>?@A"B�CDE��%&'()*�
� ��st¸H ��N��;X6����½�º����À�shallow parsing�-
���������� ���������
���
. ¸H&��� �"�Ûtãw {�bL6�½�º¸HM�#�;��ÑL6:
M\;�67� V#-���bL6�¢Á��M� /0��²} �6 &�
����÷�u�½ b�M��log-linear ÷��� max-margin-. ÌÍ�V�
Þ���ûÆÑL6M�#�¢Á��;X67/0����À-.;V��±
�6 Conditional Random Field (CRF)�&�: �¢Á����L67�
� CRF� max-marginÍ;&��½ã��L6Êà� �®¢ô��-�;÷��Ñ
�V#ML6M�£·�ÎÏ-©�_��\ f� �!#Q}�Êà� �ô���
"®#$�ï¢Á½$�½\ëm-��L6:M;��%�ÕÖ;£·��½�º\
;�6V#�L67��;W�QMU&�½ã��L6 'á Êà�(Ç�V#ML
6M�©�_��\��L) CRF ;~ÇN-���*�++L6:M&,��7
Shallow parsing �^� W�b�-�. %���L6:M;���M£·\ëmM
�# \�/;X67�
í<=½��¢;&���0����L6Óá Êà�\(Ç;���OW�Ó�Ô
;£·M�½�º\ëm-ÌÍef�Q (Tsuruoka and Tsujii 2005a)7�
�
������������ �������
� Ûtã�bL6�½�º ��&� ( )ottPn
|...1
û�ÑL6��¹M1U6:M\;
�67Q¹�n
tt ...1
&�½ã�o&ÛtãM����L6ÊàL7:�&�á V
#��åL6:M\;�67�
( ) ( )∏=
−
=
n
i
iin
otttpottP
1
11:1
...||... �
::;��½��� Ðá»�_2÷��3�L6�p!P�X6�½O>Úp4ð
5� �6�½&�� �½ �Ô�6�´µ�-�M3�L6M�á �\7>�67�
( ) ( )∏=
−
=
n
i
iin
ottpottP
1
1:1
||... �
:�\MEMM;V�(Ç5�6�8O>9: �å;X67W;<2�9O>8��
åL6:MW;�67: V#��åL6:M;�?p?p �½=>L6:M\Û
?-�@���-P�AB l¶£·ÌÍ� !!(ÇL6:M\;�67�
� CÏ-D&�8O>9: �å�Q�"�l¶£· Êà�M� &�=>b�
8E �½���O�U-�:M;X67F��9O>8: �å��Q�"&�=>
b��9E �½���O(ÇL6:M\;�-�7�
� �@L6l¶£· G�O>sU&�Êà�M� &-6H��� ��(Ç�
Q�7Ê��/�L6�½��L6��&ÃÄ�CÏ;X67: V#-IlO>�
Toutanova >&�4J V#-�_��½ã�b� 1U6:M;�K9 �½ =
>�^� LE �½��\(Ç;�6V#�� �¤t ���½�º;ûÓMÙü
Z� �67�
�� ( )∏=
+−
=
n
i
ii
otttpscore
1
11:
| �
���������� ����������� ��
���
Q¹��: ��Ñ�&��pO ��D\X67?p�&��Ô NO�º��?p
�½��\ÚC��±� �6Q}��� ?� ÚCPQ��\R� �6:M;
X67: Q}���ï FË�V� &MÙ\���SJL6:M\X67!Q�:
��^º6ûÆ�½ãT}6Q}�&�Uh�#ÎÏ\XP�ÓØ-�*�~
�L6:M\,��7�
� í34ÌÍ;&ðñ V#-�_�1U6 ;&-��LH �å +G1V
L6:M�VP�LE �½��(ÇL67WU&�oÛtO>-6 �Ô �å
+GLH 1U6MX YZYZ[ V#�-67�
�t1
(a)
t2
t3
o
t1
(b)
t2
t3
t1
(c)
t2
t3
t1
(d)
t2
t3
o
o o
�
�� ������� ��� �����������
� �
:�> +G&�4J yp ��bcL67�
( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )ottPottPotPottPd
otttPotPotPottPc
ottPottPotPottPb
ottPottPotPottPa
2321231
1323131
2132331
2312131
||||...
||||...
||||...
||||...
=
=
=
=
�
� ��\]^��bcL6+G_`L�&�Úa� Ût �½b�L6���LE
�½��\�U6:M\±O6734ÌÍ;&��å +GM�½ãLH 1U�û
W�Ô Ó��½ãM+GcdL67?e��fÛtO>-6Ûtã�b� &
'g\hi[^K �åjÍ\X6Q}�#®�¯-��çuè[;&�Ûtãk�b� l
Dm�"� ©��\ÎÏ�-6:M�-6\�89&�"®#$�ï¢Á½$�½
�#:M;�%�ÕÖ;ûÆånpºcL��çuè[ef�Q7�
� 5>��greedy-jÍM� ��Ôo Ó��½O>p�� +Gp} q�ÌÍ
wEasiest-first{34� �67Easiest-firstÌÍ;&�ûW�Ô Ó�+G\_&�6
rs&-�-6O±P���±} ÓØ-�½�º\ëm�-67�
� YZYZ[t YZYZY�34ÌÍ MÙM�½�º ØÙßL734ÌÍ�V� L
E �½��(ÇL6:M;?jäMEMMO>���MÙ\äð� �6:M\±
O67!Q�Easiest-firstÌÍ;&��M2.MÙuML:M-��Å-ØÙäð~
�� �67vbFmM� W�SVM -.(Ç�Qw �*�M§H �òO��
4ð MÙ�x5-©�_��;~�� �6:M\±O67�
���������� ���������
���
ÌÍ� MÙwy{�ØÙ�
\z{|}h~�~}]^
�}�ziz{i����z�\��z}���^� ��Z�'� ���
����ziz{i�}�z�\��z}���^� ��Z��� ��'
�}�}h�}h]���}z�{�|~� ��Z��� [����
��~�}~zi��~z� ��Z��� '�Y��
��~�}~zi���~z� ��Z[Y� '���[
����������}]z�{h��� ��Z['� Y�
�� ������� ��� ������������������������������ �������!���""#"$%�
�
ÌÍ� MÙwy{�
�}�Z��}z�{�|~�\�{�z�h{���}z���Z��'��Y^� ��Z'��
�}�]}�z�{h�\ {���h~��'��'^� ��Z[[�
¡�¢�\£�¤}h}¥��h��¢��¦�}¥��'��Y^� ��Z���
§¢¢�\���hz~��'���^� ��Z���
��~�}~zi���~z� ��Z[��
����������}]z�{h��� ��Z[��
�� ����"�� &'()*� �+������������������������� �������!���""#"$%�
�
ÌÍ� ¨�Ô� Æ"Ô� � o�
¡�¢�\©��{���h��¢�z~�¤{z{��'���^� �YZ�[ �YZ��� �YZ��
¡�¢��{z�h��\©��{��h��¢�z~�¤{z{��'��[^� �YZ�' �YZ��� �YZ�[
�}������¥}��ª�hh{��\��~�]��}�z��}~^\«��h��}z���Z��'��'^ �YZ�� �YZ��� �YZ��
�}�]}�z�{h�\ ���}��~��h��¢��¦�}¥��'��Y^� �YZ'� ��Z[�� �YZ��
��~�}~zi���~z�\¬�'��~}]{h�i{��}�^� �YZ�� �YZ��� �YZ�Y
����������}]z�{h���\¡z��z��h������~zi{��}�^� �YZ�� �YZ��� �YZ��
�� ������� ,-��.!�/ � �+������������������������� �������!���"0%�
�
����������� ������������
� �½ã�bL6l¶£·ÌÍM� &�®�¯HQV#��MEMM� CRF�V6j
Í\°��±� �67:�&��½ã�° �±\��� �½ãp�L6ð;�
�-´µW� �6O>¹M�U67�
� ���b� sliding windowj�;&��½�° ²B�±³´�����Û-
6µ��½ã �@��M� M>U67���À-. ��ï�^� &�
MEMM� CRF;X�&��±q6 �¬�½�(Ç� �í¶·Mp-\P �¸
�ï&>&> �½�¹ºL6ÎÏ\X6\�sliding windowj�;X�&��¸�
ï� !!l¶£· £·b�ML6:M\;�67�#L6M��¸�ï���±
���������� ����������� ��
���
Q6Êà�-.�»L6:M\ëm�-P�Ê�¡®m���^º6 NER V#-�
�¸�ï\k��"�&���-Êà���(ÇL6:M\;�6M�#(D\X67
: V#-�µ���0�����±Q6Êà��»;�6ÌÍM� &�
Semi-Markov CRF-.W345� �6\�©�_��\���:M-.O>�¡®m
�� NER-.;~Ç�&�±� �-�7�
Sliding windowj� ìD&��íN�&�fÛt �^� h\hi[^�' K µ���
0��DUðÉ6ÎÏ\X6Q}��£· Q} ©�_��\���-PL)6M
�#D�X67�:;89&�Naïve Bayes �@¼2«���M� (Ç� ���
½¾�QW ¹º£·b�ML6ÌÍ$%Q7�
� YZYZ��34ÌÍ¡®m�� ���À�Æc�Q¿qßL7�� ÀÁ
�u��«ï�Â�� ûÓFmüZ� �6�*��&Ã6W �Ó�Fm;
���À\ëm;X6:M\±O67!Q�34ÌÍ CFG Þ�¬�½�ÆÇ�
Q¿q�ÃÄ�ÓØ-CFGÞ�Ä++L6:M\ëmM-�QwÐ ÅÆ[�$uÇ�
� o ��{�\�~���{|���h���~����'����^7�
�
� ¨�Ô� Æ"Ô� � o�
¡���{�z��}]z{��¢�]��h}�É�§¢¢�\«�{��'���^ ��Z�� ��Z�� �'Z��
34ÌÍ� �'Z�� ��Z�� ��Z��
¢�¢¢�\��h|}��}z���Z��'���^� �[Z�� ��Z�� ��Z[�
���\¡}zz�}~Z��'���^� ��ZY� ��ZY� ��Z��
�� ����$�� 123456�789� ��
��������
(Tsuruoka and Tsujii, 2005a) Tsuruoka, Yoshimasa and Jun'ichi Tsujii. (2005). Bidirectional Inference
with the Easiest-First Strategy for Tagging Sequence Data. In the Proceedings of HLT/EMNLP 2005.
(to appear)
(Tsuruoka and Tsujii 2005b) Tsuruoka, Yoshimasa and Jun'ichi Tsujii. (2005). Chunk Parsing Revisited.
In the Proceedings of the 9th International Workshop on Parsing Technologies (IWPT 2005). (to appear)
���������� !"#$%&��
l¶£·\ÃÄ�Å °�cÇ&ÊË :M;X6\�~� ���ÆÇL6�ÌQ
� &�ÌÍ�� ÊÎFÏ��1V�QÌÍ efM�l¶£· ¿q���Ð
Ç;�6��çuè[ �G\êëìM-67st¸H�����¸H ÊÎFÑP
Ò2¹l¶£·ÌÍ <=&�í¢Á¬îï� HI; Ó ?p;XP�� Zq&�
Ã!"ôõ�}6�ö� åæM�Ô÷��M !"M�#ZqXÉ �67�
íz;¯HQÚp ÌÍ&����½�º�ÔöÕ�ÔÖ� ��h|}� -.�st¸H
°�×Ø ��å�Q} ��çuè[M�Ô÷��M !"�#W ;X67
: Úp ÌÍ&�iã�½�º���ÙÚN-��¹º;-��BN-��ÑPÒ
Û:MÜXL6D;�ÝØ�°!� �6�ÕÞM�ßjäFàp\�©�N�&
&6O�áâ;XP�~Çð�ÃÄ���-ÌÍM-� �67~���89\ef�
���������� ���������
���
Q £��¬ã�*�\äåæ£�� �����bL6���½�º^�ÔöÕ�çÞ�è
&��ÕÞM�ò-FmVPÓØ�~�� �6:M��� �67�
����¸H�st¸H ��;�iã�½�º��M� ��Ñ;�6��&ÃÄ
���789 Zq&�:�!;�Ké ��M� �K9��ÔÑMÓMÙÑ�l
� �Q?@ ���? êÝ%;%6:M�Op�¡z�z}i{�iã�z~ MÙ&6O
��ÔN�~�L6W ;����5!ë!-���ÆÇ5�6:M\��;�67�
�
'()� *+,-./�01�2/3�45�
� � � ��6789:� 01;<��/=>�
�?�����@ABC����
st¸H�����¸H <=i!N��#Q}�&�ìð íIN-HI;&-
���~ st�Ç ~î�ï�Q�Â�[��¢N-jÍI\êëìM-67oxÐ
z; _�Þ����� Í÷�� ef&�HIO> ��¢"Q�N-÷��ÑM
���O> Â�[��¢N-ÌÍ¿×pºQW ;X�Q7Q¹�Â�[��¢N
-jÍI���ÆÇL6Q}�&������ !!²}Qðä_�Þ�ñ¹º;
-��HIO>òäL65!ë!-��w���+ ó-.{�ô��HIO> ò
äõö\~î �����. V#���6On6ÎÏ\X67�
M���m��Á¬��Ü÷M��Q���øù;�-�W MstM¿×pº6
�&����+ óM��Qst£ HIO> òäõö¹º;-���� ËÀwm
��Á¬�{O>òäL6��W�ô�-º�&->-�7: V#-ð�½��_�
Þ�ñ�X6�&�ð�úp�_�Þ�ñ&�l¶£·¦�� st¸Hwoxozøù{
fü�û� �� CÏF\Ó!� �67�
í¢Á¬îï�;&�m��Á¬���QËÀ¸HM����¸H ü"�lL
G�O>�m��Á¬�<=: �ý\Ó��Op�~� ��; ËN����¸H
: ÏTWÓ�äåæ£���ýDX� �����: m��Á¬���þ}Ò
ÛY���Q7� Zq&�äåæ£��; m��Á¬�\�öÔ'*m��Á¬
�^M������Ü÷�������p�_�Þ�w�öÔ'*_�Þ�{M� �
e5� �67!Q�äåæ£�� ����&�½Ä st�ÇM&�-6��w�
D ��ÇtM�t��ú-�ã.�D��Ñ£� K�-.{\��%>�6:M�
+ +G !©N-P-. H�O>����½�+ ó-.�st�� �ôY�
��Q7:�> ¿q&�oxÐ�oxo�ox�z <=��;ÐÇ5�Q¹º
;-���� <=½��¢wÚy��{ ½��¢�V� �±� �674J;&�
_�Þ� õÏM �¦� �ú�ô�p� ¯H67�
�
D�EFGHI J/K3�LM�
±�±�&�l¶£·¦�� ��st¸H ÌÍcÇL6Q} £·^V×gs
���M� �MEDLINE���¦��ð �¯��ï����Çt���YÇ®¦
������+ ó�½�º�Q_�Þ�wGENIA _�Þ�{ef�Q7��-
���������� ����������� ��
���
. �����b� L;�~Çëm-FmßL���º�p� W�: �� �
����b� &�� ��\X6:MO>�±�±�&MEDLINE�¯��ï��
b� 1U>�6�P ����»�ï��¢� �:#M�#j�;_�Þ�
YZL67�? ����� � ���½�ºL6:M�V� ����Ö �±�
QMU&Çt�c Q}�.2-stËÀ\Ì\OPM-P#6O�øùL6 ��
G 6:MW�NM� �67�
b�ML6����&�MEDLINE ���¦��O> humanwÀ�{� transcription
factorsw !"#{� blood cellsw$%&'{ Yp��(�)M� gh5�Q¿q
�¯��ï�*;X67:�&�äåæ£�� �;WO-P+�<=AB�-�
�6Q}�,-¦�� �¢Á���(ÇL6�&� P¡��¬&Ï�;&-�7
�O��AB°�.��L)6M���Çt�Ê�/Ë-\�0�±QP�l¶£·
¦�� �¢Á�����-Êà1U6:M\;�-�-67?j�äåæ£��Ê
� st£N-Êà&�: 2PÒ2¹�¯��ï�*O>WÏ��3>U6:M\
;�6M1U�: AB �¯��ï�*Ç� �67�
������������� �������� �������� �� ������ �
��!"�#$%&'�(�)*+ ���!"��,�&'� -�� �������
� .� ��(/0�1234!"�/563�789:;�<=>�?@A�/0�12
34!"�/>B3CDEFG4�Institute for Infocomm Research�HAIJ��9K2
E LMNOP QRS���� GENIA!"��%T�UV� ��������9WXYZ [\
�]^�_WYZ(/0�123CDEF`a��bc�0�12�def0�ghfK
ij0�12�k9lm4nop/qrEFs
s
��EFGHI NOP1J/K3�QRSTU*+,-./�
GENIA ���!"�(Kim et al, 2003)��:;#t�u� v�(�\�wx/$
yzE##{9�|c���9}DC0A"�;\f~�\]^�����2/zEFs
: ï��º Q}��89& GENIA m��Á¬�M45x67-�@�i(X
3.4.1)++�Q7GENIA m��Á¬�& Substance(/Ë-)�SourcewÚ8{^V×
Other(� �O^M�#è¯_�`�a��O>-6õö 9ç�± ó+G;���Ç
t�&: :��) �r�OÜ÷ï�M� �T� �67_�Þ� .©;9
;&�:��)¹º;-��Ü÷ï�M� �r� ��)_`� WV��!Q
<D Ü÷ï��TL6:MW1V�K� �Q7�O�~��½�ºY�=}
%6M�Y�>�Ï�-?@ËÀ\X�&�:��)APÌ 6:M\ëm;X6
:M\±O�Q75>����Çt�&Ì��»tWB8���? �;<D Ü÷
ï���L6ëmF\X6\���B� c�CÚ;&�� D�²B� ?p
Ü÷ï�APÌ 6:M\ëm;X�Q7Q¹���8 GENIAm��Á¬�;
&�/Ë Ñ£N-FË %���� �@� ^P�lm��L6�@�� �-
�7���ä��; �A-. /Ë lm��L6õö�»���8�½�º5�
�6Çt�b� 5>��FEô� ��=�;X67
���������� ���������
���
�
�
V�EFGHI WXJ/K3�
MEDLINE�¯��ï��b� �FÇ ���*�;X6 Junk TaggerÆÇL6
M�Wall Street Journal�b� 96.8%X�QMÙ\ 83.5%!;u;67�:;±�±�
&��Çt_�Þ�M�ß�¯��ï�G���b� ���½�T�Q_�Þ
�YZ�Q(Tateisi et al, 2004)7��_�Þ�;&�¯��ï� Ût� Penn
Treebank(H���¡¼��£;YZ5�°��±� �6���+ �½�_�Þ�)
G��������APÌ 67�½�T&�!rI¸HM� Penn Treebank
��ï#®Ä�Ç� l¶N���ï���A�Junk Tagger;J��½�º��K
Ì;LML6�M�#ÌNM�Q7::;�äåæ£�� I �Ê�MO±�6Ú
p ��\äßQ7
Ðp&�/Ë-��½Ä ����->&PG ��ï�M%-5�6_�»�QR
-. ñó\S�!�6:M\X6(WT2,3,7,8-tetrachlorodibenzo-p-dioxin):M;XP�
W#Ðp&�LM�#�����\��Çt� t ���º�U#:M\��:M
;X�Q7Ê��: ��;&K--. �±q6�-�;&-�/Ë-�&'V-
\� Ü;W!6\WT/Ë- Jun�&'V- Jurkat)ÊFà;��-�MX½-�
Yé�U#:M\�O�Q7�O��:�> ��D&��Çt�²�� ^P��
M2.& D�²Bér���p}6:M\;�67: :MO>�Ð{��Çt_
�Þ�O>��ÇtM��5�Qµ�ZPc� ��> t+ZL6 Ût�K
Ì;��pºQ,-YP�Ú{��WM���ï��ALML6MMW���
Çt+ZL6 ��ï��X>Oß}l¶N���APÌ �o{Junk Tagger&�
�½M[\�-�V#�]P µ�����½pº�y{KÌ;LML6��&X
>Oß}l¶N�APÌ Q��&M��W ML6�ML�&^��Q���º\�
U67!Q��-�MX½-� Yé&+ åæ-.���½(ÇL6¢ÁG��
s
s
s
s
s
s
s
s
s
s
s
s
s
s
s
s
s
�: ��$���� ;<=>? 3@ABC*+DEFGHI�JK%�
+-name-+-source-+-natural-+-organism-+-multi-cell organism
| | | +-mono-cell organism
| | | +-virus
| | +-tissue
| | +-cell type
| | +-sub-location of cells
| +-artificial-+-cell line
|
+-substance-+-compound-+-organic-+-amino-+-protein-+-protein family or group
| | +-protein complex
| | +-individual protein molecule
| | +-subunit of protein complex
| | +-substructure of protein
| | +-domain or region of protein
| +-peptide
| +-amino acid monomer
|
+-nucleic-+-DNA-+-DNA family or group
| +-individual DNA molecule
| +-domain or region of DNA
|
+-RNA-+-RNA family or group
+-individual RNA molecule
+-domain or region of RNA
���������� ����������� ��
���
&X!PCÏ;&-��&O��_Y� :2>2L6VPW^�� b`;�6:
M �#\CÏ;X6M1U�I a>--. K-�<=l�--. B�X½
-�M� Ò#V#� Penn Treebank ���º�_LML6:M��Q7
��_�Þ�& Penn Treebank�_b�Q��M���Çt�½ p�Q_�Þ���
��½ ��»�¬�Q XML�� 'p ��;�e�Q7�> ��;&���
Çt&?p4ð ��ï�O>-6W M3��Q\�I �¯��ï�;&�>�
Q ÜD;I õÏcd�V#ML6:MO>�òe�f gÙ\Ó��Ê���
�Çt�p� ��ï�VPW5>�+�Ûe;¥½L6�h,��i,-.¥�L
6t�° òe�f\X6wWThomo- or heterodimers{7: V#-�"���Çt�
½;��ï�\�A5�6:M\R:67: V#��A5�Q��ï��b� &�
� ¹±P�jkjAPÌ �6wX YZ�Z'{7�
�
�
�
�
�
�
�
�
�
�
Y�EFGHI 4Z[J/K3�
+ ó_�Þ�\Tateisi et al, 2005)&��Çt_�Þ� b��¯��ï� è¯
G���b� �GDA-DTD lm�Q DTD Ç� PennTreebank M�ú +G
XML��;�T�QW ;X6wX 3.4.3{7
�½�º �_&�n Penn Treebank���� �6\äåæ£��;&��Çt\
Õ�&Io�.pÛ�"\X6-.k�w�D ÛtO>-6{�"\�����Ç
q¡rq��r¬h�q��rz�}���}~}hz����}��q���rq���r��q�
��i¡�s���tu���urq��rz�}���h��h��q���rq��r{��q��r��v['�¬wi���}�}�����{~z}�{h}�
�}����z��}�q���rq���rq��rz{�q��rq��r���~¤��¤}¤���h}���]�����]z�{h~�q���rq��r{��
§¢�� q���rq���rq���rq���i¡�srq��r��~� q��r~z���}�� q��� �x��tu���u�
�}�tu���u�rq���r�
q���rZq�¡r�
�
\¡�\���¬h�\���z�}���}~}hz����}�^^��\��i¡�si���\���z�}���h��h�^�\���{��\�����
v['�¬wi���}�}�����{~z}�{h}��}����z��}^^�\���z{�\���\������~¤��¤}¤���h}���]��
���]z�{h~^�\���{��§¢�^^^^�\�����~�\���~z���}��ki��^^Z^�
�: ��$����;<=>? LMNO*PQRST UVW XYRZT [����\���]��. XY�
����������s
�s
��M��M��M���P��s�M���������MN���MN����M�����P���M����P��s�M�����P��M�����M��s
��P��s�M��������s�M�������P�M����P�M���M��s
��s�����������������P�����s������������������ss
��P��s�M������M���s�M������M���� �M����s���¡¡�������ss
��s���¡¡���� �������P���s��s���¡¡����� ��������������P����s
�s
�����������s
�: ��$�"� ;<=>? &'O*PQ
���������� ���������
���
t�µ +G.:!;�X6�&. V#��æL6Op}6 &y,;X6M=
,5�Q7���º �"M�ú���Çt�µ +G& D�²B�-�W M1U
>�67�Q\� ���Çt��� & DO>PG��æ�Qj\��ßÇt�b
� �æ\q�6z{\-�-� V�Mb`��+ ó�½�º �&��Çt�
+G&�æ�-�:MM�Q7�½�º&|��KÌ;���������&�½�
º�U�QD�p� ;�6¹º_���� ^� W>#V#��Q7YZ�Q }¢�
2~®�&��IN� �}hh���}}��h| ���¹º��Lj ���e� �67�
Y�½� �òe�f +G�º\y,;X6:M\±O�Q7��_�Þ� %
;W¯HQV#��I �¯��ï�;&?µ���Qòe�f\��7� V#
-���âL6�& + %->rÜ÷��L6Ï�-Hå\ÎÏ;X67QM
U&X YZ�Z�&�ß.�¬�i[��h��¬�i[�i¤}���z}����h]z�{hj�bL6Úp +GÑ
ßL7�
�
�����
��������� ������
����������������������������������������
�������
��������������������������������������
� �!��"���
������
� � � � ��� � � �
� � � � � � �
� � ã�s�� � � ��h]z�{h �
� � � � � � �
������ã�s�� �h�� ã�s�� � � �
� � � � � � � � �
�����¬�i[��k� � � ¬�i[��¤}���z}� � �
�: ��$�$#���^_`a�b>W#����c�>W#�d#e�c!���c�f����!��g��Lhi�
jkTlmn�be�c!���cgopq�r�stuvwx��yz�qR�
YZ�Z��i� &ð¬�i[�¤}���z}��Op ¬�i[��¤}���z}��;X6V#- ��h]z�{hñM�#
åú;XP�X YZ�Z�i�&ð¬�i[�¤}���z}����h]z�{h�^V× ¬�i[��¤}���z}����h]z�{hñ
M�#åú��ÌL67: .;>\M��O b`&�st£��ML6����
��&pº�>�7F��ä/£ ����&+ +G��L6ËÀ\ê�� �6Q
}�+G�º�� Y�\y,-W M-67���: V#-�" +GÑ��O�
��� ËÀÑPÒÛO�QMU&�Ü÷\±O>-�Q}+G\X�!��-6µ
� %3ß� X�!�Få�L6 �ÎÏ-ûS� +G�����T�
W>�� åú�C�Q�� +Gst£>\�TL6�-. �d�ô�GL
6:M\�äåæ£ %->rÓÙ���Ñ5�Q��L���� �½�º Q
}�ÎÏM-<#7�
���������� ����������� ��
���
������� ������
������������������������������������������
�������������������#�#��$�����%&���������
�����
�����������������������������#�#��$�����%&���������
���������%&��� �!��"��������
������
� � � ������ � � � �
� � � � � � �
� ��� ãh�� ��� � �
� � � � � � � � �
�������ã�s�� � k'� ¬�i[��
¤}����z}��
��� �
� � � � � � � � � �
����¬�i[�k� � � � � � ��h]z�{h'��
�: ��$�$#]��: ��$�$#� {|}~oq���Lhi�
jkTlmn�be�c!���cgopq�r�stuvwx��yz�qR�
j"0 kT�bf����!��g{opq�lmnz�qR�
\�EFGHI ]^_P`a+,J/K3�
��YÇ®¦��_�Þ�&� �;ä�� ��YÇ®¦��\. V#��5
� �6 OdßN�ñ¯L6:M�NM����Çt_�Þ� b��¯��
ï� è¯G��;X6 ����¯��ï��b� �Caderige¢Á¬îï�w2�
�{� �»5�Q DTDlm�Q GENIA-event DTDÇ� }¢� ��;�T�QW
;X6wX YZ�Z�{7�
�
�
<sentence id="U93136418:7" title="no">
<genic-interaction regulation="inhibit"
type="protein-expression">
Finally, <exf>using <ex>electrophoretic mobility shift
assays</ex></exf>, evidence was obtained that <gaf1>
<ga1 role="modulate" type="protein">IL-4</ga1></gaf1>
<if><i>inhibits</i></if> <gtf1>LPS-induced expression of
<gt1 type="protein">AP-1</gt1>
protein</gtf1>.</genic-interaction>
<non-genic-agent-interaction type="protein-expression">Finally,
<exf>using <ex>electrophoretic mobility shift assays</ex></exf>,
evidence was obtained that IL-4 inhibits <ngaf1>
<nga1
type="exogenous">LPS</nga1></ngaf1>-<if><i>induced</i></if>
<gtf1>expression of <gt1 type="protein">AP-1</gt1>
protein</gtf1>.</non-genic-agent-interaction>
</sentence> s�: ��$���� ;<=>? ���42�@AO*PQ�
���������� ���������
���
��YÇ®¦���½& �Ûe;�T5�6W M�� �&��YÇñ¯L
6W M�-�W ��@5�67��YÇñ¯L6 �&�“Assertion”�“Regulation”�
“Type”�“Uncertainty” �“Self-contained” �“Confidence” ��@ �Fà;���
O�� O �F���YÇ FËM�@�� �~Ù��� |¿F�Y�> �
�Ù�L6:M\;�67��B� ��YÇ&���YÇ>wAgent{��NwTarget{�
YÇwInteraction{O>+Z5��w��~5�Õ��ÕÖ��Ú�~�ÌÍ��L6
�-.����M� �TL6:M\;�67�
��������
(Kim et al., 2003) Kim, Jin-Dong, Tomoko Ohta, Yuka Teteisi and Jun'ichi Tsujii. (2003). GENIA corpus
- a semantically annotated corpus for bio-textmining. Bioinformatics. 19(suppl. 1). pp. i180-i182.
Oxford University Press.
(Kim et al., 2004) Kim, Jin-Dong, Tomoko Ohta, Yoshimasa Tsuruoka, Yuka Tateisi and Nigel Collier.
(2004). Introduction to the Bio-Entity Recognition Task at JNLPBA. In the Proceedings of the
International Workshop on Natural Language Processing in Biomedicine and its Applications
(JNLPBA-04). pp. 70--75.
(Tateisi and Tsujii, 2004) Tateisi, Yuka and Jun'ichi Tsujii. (2004). Part-of-Speech Annotation of
Biology Research Abstracts. In the Proceedings of 4th International Conference on Language Resource
and Evaluation (LREC2004). IV. pp. 1267-1270.
(Tateisi et al., 2005) Tateisi, Yuka, Akane Yakushiji, Tomoko Ohta, and Jun’ichi Tsujii (2005) Syntax
Annotation for the GENIA corpus. To be presented at IJCNLP 2005, Jeju, Korea, October10-15.
���������� !"#$%&��
�8���Çt_�Þ����_�Þ�& '��� �¯��ï�wõ|��t{�
e��]µ �� <=>�VPÐÇ5� �67�
äåæ£�� ����¸H���[ ef�^� &��!¹��� �"K9
ef>\�P� x67-_���� ���[ ef����� �6:M
O>�<= �����§¨\�} y,�-� �67GENIA _�Þ�&�e4¶
: �� _�Þ�M� û�67 W M� ��Q-ÌÍef Q} _�Þ�M
� ¹º;-��<=Zq §¨�#Q} �_N-_�Þ� �e�7� �67�
� �_Á)�£ ½��¢& 20050 ISMB�. Biolink(�ï���¢�^º
6�-äåæ£I _�Þ� �Ç/�����GENIA _�Þ�\ 21 OMww10
O��{�§H ��N���:M�þ� �67�8�� �2����Qî�
��-. �� ½��¢\äåæ£��; _�Þ�YPeW� �6\��öÔ
'*_�Þ��.��N� �¦� �ú�ô� �6W &-�7!Q�::;�
þ�QW 4]����*`�� ½��¢wInforcomm{\¥øù�± �ô�)®
� ½��¢wEML{\�öÔ'*m��Á¬� lm�#-.�GENIA _�Þ�
����Q ��dWW!� �67:�> $%M����� ef�� �
6H���¦¼��£��ÔÕ�M K� � p}Q¥�<=WW!� ^P��
öÔ'*_�Þ�M� �¡\�� äåæ£��; ����¸H�TU6´µ&
�} ���7
���������� ����������� ��
���
'(b� cdef�ghij�kl�2m,no�pq�_��
� � � � 6789:� 2m,no���/=>�
�?����@ABCrs���
v2�Qî�½��¢&��67-����bL6���-©��ÎÏML6st
¸H¢£L6:M¤¥M�Q½��¢;X67í½��¢&��~�L6Q}��
�¦§¨�^º6�ã¸HáâOp�ÔV��#j��Qî¯O> �67����
��� ÓØ�©ª-«²M¬�-.�p� <=��Q7�
�-ZqM� �(a) �¦§¨;�®ÄY��©�l¯HO>��ã¸H!;�Ô
N��#Q} §¨���� .©M~°�(b) st¸H½��¢M ���V6��
� MEDLINE �¯��ï� HPSG + åæ((a) Zq(Ç)�(c) Qî¯O>Ó
Ø�H�¬"Q�Á�)L6±�¬� .©M~°���Ç�Q 2 ²H�¬/�³
Õ 1.3TB´Ù Qî¯��� «²�(d) st¸H½��¢M ���V6��� ®
ít CFG+ åæ((a)(c) Zq(Ç)�\X67�
:�>�VP���M� ��� Qî¯_�Þ�Ç���67-©�µ¶·K
�Qst¸H�#��v2�Qî�\++5�Q7\�^�\�^&�!"~�\ox¸z^
;�þL6:MM��íz;&�\�^��\]^�p� �þL67�
�
D��tuvw�xyz{<|}~�;<�w�&������/�Q���
�8�ÓFm©�§¨M� �DÏSO>D¹S ©�l [��¢��~tD £��~ ï
� Á�P�±u�a��(�ï;¿"�Qï��©�l\� M-� �67!Q�
� V#-ï��\�¸�Þ��X6�&°B�<DB8� �6�¦§¨W?e
Ñ� �67:�>&�8��� Ǻ�b� �ûÓ »ÇbFm§WQ>L©�
l§¨;X67�O�� ?j;�:�> §¨�ã¸H ���4] ¼�Ä\á
â�(ÇL6Q} v2�Qî�§¨&½� �-�7QMU&�¦§¨;&�2~®
� ¥��¾¿�¦-.�Û?���[§¨�ï��§¨;�\µ�N�&^mH��
�«�½���[�V� �IN�3n5� �Qlm\ìu� �67: :M&�
Ê��¦§¨�^� �¢Á½[ ef��ãÑ�®�����-. _��a�
�À�5é67K9 ��DåpL6v2�Qî�&�D345� �6\�F��
�> v2�Qî�JK�ÁàL6Q} _��\��;X�QP�̄ H>Â�Î
ÏML6Q}�LH ©�l�¥½L6§¨3n;�-��"\��7:�>�V
� �� Ã8N(Ç>\�¦§¨O>ÄëO� �6 \�/;X67�
89\.©�~°�Q��� £}�&�½Ä �î����Åk�Q®��2î��
;��O��D ©�l�Õ��#:M�í�®��M�Q���;X6\X
YZ�Z[^7X �; ÆÜ £}�v�����������w\�î�M�ú ¢Á�¢�;�::�f
� _»�)KdL67DÜ ���&��8 ���S ¢ÁG�è\M��&���B
� '¢ÁG�èàp 'Y�Ç��^�Õ�ÈY� �6:Mß� �67�
�e hostname”&·K5�Q_»�);�:��VP�Ç��;?É�_»�)
hostname \~�5��Ê�f�����¢ÁG�O> cd\¿"5� ß5�67
���������� ���������
���
~�&ÃÄ�ËØ;�ðñ §¨;�RIO>ÌÍ!; �Z� Ç4J;�#\�¯^7�
Ç��&ÎÏ�£íÐ�¸�Þ�M�Ñ�¸�Þ��.o5� �6\X YZ�Z'^7�
�
����������� �
�������������������: �������;U[ O�@��2@2@(��*Q�
�
����� �
���������@PQ��Q(� � � � � � � � � � ����@PQ��Q(�
�������������������: ����"��;U[ �QAO@��*(�
�
GXPM@Ò ���M� �dsh� pdsh-. �ssh� rshÇ�QÄÓ_»�)RI
�ã��#����mpd � pvm -. ÓØ-�ã¢Á½[��¸�\X67�
�>M GXP&�4J V#-D�� D;�-67
[Z �DÇ���bL6ÓØ-¢ÁG� RIT
GXP &AB ���M§H �SD\��(100 S�U6)§¨;�?ÔO>ÕÔ
ÓØ;X67���&�AB ��� ��(dsh, pdsh)&_»�)~� Q×��s
p}Qu÷��Á½®�~�� �6 �b��GXP&��Q2äZ5�Q�
fÁà� �4Ö _»�)~�ÓØÑ� �6:M\���×T� �67
ÓØ-�ã¢Á½[��¸�;X6mpd&GXPM�ú +G� �6Q}�
���������� ����������� ��
���
dsh� pdshVPWÓØ;X6\�¢ÁG� RI Q} ½��u�½+GÇ�
�67?j GXP&ó+GÇ� �6Q}�QMU& 150S�U6V#-��
-SD;&�GXP & mpd VP?Ô´ÙØ�7X YZ�ZY &��> ��� Fm§
¨ß� �67í¢ÁG�~�L6 �OO6ÕÖ§¨� ^P�J:��
�.ÓØ;X67�
�
s
s
s
'Z JK�Ø�L6:MûJO>öh�o�Q.©Ù��
GXP&<D �2~®����[¥�� �-�§¨;�ûx� J�.�;I
�:M&ß}O>.© ���� �67ûJ� GXPRIL6?Ç��:®�
����L6M�� �O ��):&�IN� GXP\®�����5�67!Q�
ï����B8L6Ç��- Ñ7-.W�IN��±�67��M� �J}
�#¼�Ä\D� #;��u÷��Á½®�ëm-D¹ ��) �Õ(Ç
eW;�6V#�.©5� �67AB �M2. ���&LH ��);�
� _»�)\�?Þ��®�����5� �6:MI3NO�� �67!
Q�_»�)~�L6Ç��LH ãÈL6:MW¼�Ä ÌÖM-� �67�
YZ �¦§¨ �úF�<Ú-a��(�ï.�1V�Q.©T �
GXP&<D ¯H)�®�(LAN)�!Q\�Q��ú�OpM:<�V� a��
(�ï�f\ô�5�Q§¨;IYL6V#��ûJO>.©5� �67AB
�� ���&�¼�Ä-\LH ��);�ß;X6:M�. ��)bÖ;
W��Á½®�\ëm;X6:M-.�~ËN��LH ©�l\?p ¯H)
�®�(LAN)��X6:M3�� �67!Q�GXP&®������¯H>Â
�WÎÏ-��(Ç>\���\�PQ��àp©�lLH (ÇL6Q}
ÛÜ�dÝ-�L6V#�.©5� �67�
�: ���������q����K��*�y{�����R�
l�B�Qz ¡q��¢¢�£¤�¥¦§¨_T©yzª«�¬¦yz®x¯°±
²r�R³´Gª«��µd ¶*�y;U[ ��B�Q ¡£¤�/·���yT¸¹ºG e�c�
�e�c���y{��±xr»¼¹º½¾¿ÀoÁFGÂQÃ@QzÄűx��Rs
0
2
4
6
8
10
12
14
16
18
1 2 4 8 16 21 42 63 84 126 168
gxp e mpdrun dsh -r ssh -c gxp explore mpdboot
���������� ���������
���
£}� & �zz�Ù�����Z�{�{~Z�]Z�Z�iz{|�{Z�]ZÈ����{}h�Þ��Þ�ߦ��]|ߤ�hßÈ�Z~�z¤� �
�e� �67X YZ�Z�&��¦§¨: �ïG�VP�Ø�L6 £x¬;X67�
�
� � � � � � � �
�: ����$��;U[�;Æ>RÇ�QA�ÈÉ�,[Æ Ê4ËsÀ1*{̽Ívwx��y�
�
����� ��� �/����kl����n+-/��+.+�
WWW ���(ÇL6<=&D���±� �6\�� ��& WWW ð;à
I� �6gh±�¬�:ghÏT�� ���KÌ� �67: jÍ;��~
N-ÕÖ�;KÌ;�6����&��\XP�!Q��gh��c �N�Ç�6�
�çuè[��±-��Ä���(�)�V� Ñ7�Q���� 2Ò%�±-�
&->-�-.�ô�W��7� Q}�WWWO> �67-���c�_�Þ�
YZ-. Q}�&��d;H�¬ "Q�Á�)�hÊ�º�#ÎÏ\X67W
;<2AB gh±�¬�-.;W�ú l+~°� �6\��������ï
G�;�6V#��e5� �6±º;&-�7!Q�WWW ����&á�;X6
Q}��W�W. V#-����ý�«²LH�OW��NML6<=�V� �
-P�©ª;ÓØ-«²±�¬�àp:M Ü»&���7�
� í<=;&�Qî¯O> ���c¢£L6Q} ���M� �ÓØ;^�-«
²±�¬�.©�~°�Q7<=�ÖIâ�ã?ä ~°����â�¨.©M¨
~°��Q789 .©�Q±�¬�&Û?-_»�)®�®��2î��à;�
�_Kd�TU>�Q x��àpH�¬Nê�;«²��¿q�H�¬åM�YZ
PK� �_cd�cLM�#W ;X67:��7QH�¬ åæ�#¢Á½
[��online�KdH�¬©�L6¢Á½[MÞ®¢;¿"�QP�GXPÇ�
�ã�RIL6-.� �©ª-«²¼\�Ø�++ëm;X67«²±�¬�&�
±q6 WWW ð ±�0��(Ê� è�¡�²�� ¾¿Oº-��-.^æ6
ÎÏ\XP�ÓØ-«² Q}�&�D\[����/�^ _aï���Áà� �D
Ç��O>Ý�rp��� "Q�Á�)�#ÎÏ\XP�89 ¢Á½[&:�
���������� ����������� ��
���
�� �[S Ç��; ���H�¬�Ç�'�S; [�����H�¬�Ç ØÙüZ�
�67Û?©�;]çL6M�[®� �è�H�¬�. «²md\X67Q¹�:�
&éÕÖ ñê;XP���kÕÖ ëÇ�#MMW��«²�QH�¬O> u�
ï�c�hÊ�º-.p}Qì"N-��M�ÇF gs�� ��ÎÏ\X67�
<=�Ö Iâ��[ íî´Ù;õ 'Z� ²H�¬´Ù«²�Q\�³Õ 1.3TB)7:
���&�st½��¢(ïð½��¢) ñò�óô>�V� �Qî¯_�Þ�O
> õ2��[ £· Q} ���M� Ç�>�Q72.5 ²H�¬&���«²
�Qï��� 4��)O>-62~®�è�¡�!Q\P���ð¯�QíÐ�
Ñ�.o5�Qï�� õ 350CPU;¸H� �67 H�¬�b� �2~®�è
�¡O>ï��: ��� ö�st�À(®ítH�¬ �c), JUMAN�V6�î
Eåæ�KNP�V6+ åæ~�� �67�ã¸H�&�&PGXPÇ� ^P�
õ�;�1,000CPU®´Ù ¸H�YZ�®´Ù;¸H� �67�
�
���������� !"#$%&��
®9ä÷�¬5� �� WWW ð á�-�����b� ���è�!"5�
Q�� ©�µ¶Ç� �~��ø���st¸H�#:M\;�6:M~s�
Q:M�Ü»\X6789 ù��V�&�:�> ¸H~�L6Q}��KN_�
� Ó�$)�Qî��v2�Qî�®�2 ½ú�̄ H> �d&�M2.ÎÏ\
-���±q6LAN�p-\�Q½Ä .� Unix©�l a��(�ï;û�P67
í<=%�;YZ5�Q��Ö&�ox¸z !"~���±��̧ H;�6���
��ÚÔ4ðäð5é6:M�Zü���� �67����¸H: �����
G�Q7
st¸H�ý ~�;89&�¾¿�¦ Q} ���[���¦§¨ Q} 2~
®����[-. Êé-$)�Qî�&Ç� �-�\���> lm& GXP \C
±P3n� �6O�GXP\X�&(�ý ~� ×Ø�;&)êÏM�U6W ;X
�Q7kÕÖ�±Q6¬�¯ ~�&�89 ¬�¯ þ®ÙûS��Q/î;��
O ¼�ÄM½Ä½Pµ¶¥��Q/î;�±��89\©�lO>w ¼�Ä�
��Q±º;&-�:MWCÏ;X67:�&��ã©�\<= ��;X� ��N
;&-�<=>�&��;XP���-��)Qî��ôU ���-KN_��;�
í� ~��ÕÖA�:Mëm�����-©�md®ÄN��Ç;�6§¨
++;�6:Mß�Q7: Zq&�����¸H�st¸H��>r�°�æ£�
� <=; ©�l§¨++L6 ���-Ì;M-67
st¸H&������-,-�m��Á¬�����ËÀ¸HMst¸HM!"
L6�VP��-©�mdÎÏML6<=�=Û:M&Î;X67�ý <=;+
+�Q�� ð���¦§¨J; 2~®����[++L6��efL6:M;�
á�C �����st¸H Q} v2�Qî���\++;�6M1U �67
���������� ���������
���
'(�� �������k�/.o+,��@��`+���� +¡��
68¢9:� �/.o+,��/=>�
�?����@ABCrs���
� í½��¢;&�¢£¤¥¦9KDC§¨/�D4©ª�«�9i¬C®f¯°±]
?A0��H²A/³´¢£¤°µ¶RSA��6·/³D�¢£¤9K2E®¸�¹
º»¼½¾¿9ÀÁzEÂÃ/ÄÅ6·zEFs
M���ÁÂ��_$Á¼0���� Q} ��«�M� eo�º�KÖMÁ
Â�� �N-®��ï����V� ���N-_$Á¼0����MÜO�½
�#�� ef�l�Q7:�&�����M�#st ��V6�� ½�
��L6st¸H½��¢M<=M��N-´DàpW ;��¶�: Úp ��
�ü÷�)!"L6Q} êÝ% ��M-67<=ZqM� �Ê�Ò%ôHM�
Ô�I��QKÖ�K�/®��ï���j�M�#�íN-êÝ%34���
��F~°���[;~s�Q7Ê�Ò%ôH��Q®��ï���j�&�
K9 ¬î��¸ Ü÷�º=}�±-�����Âu�ï¬î��¸�V6®�
�ï������®Âº ~��QW ;X67
D�£¤¥¦§<�¨%`+���� +�©�ªP�
Ê�Ò%M&�/�uè[;�I� �6���[ ÈI\��L6��;X67Ê
�Ò%&�K _$Á¼0����;WCÏ-�AqQ� �6M1U>�67�\
&r2;�6MË>rË>r�ßuè[;��X� �6M�# &� ?W;XP�
��ïÁ¼�wð��ñ{M� Ë>� �67
Ê�Ò%ôHÇ�6:M�V� �=}��Âu�ï¬î��¸�V6®��ï
��� ¢Á�_��} ^O-� LÛ����Âu�ï¬î��¸�V6®�
���j� ~�\��;�67
í<=;&�Ê�Ò%ôH�� ®��ï����øô� �6��>Ö ��
���½�#���[ef�Q7: ���[&���M¹�M�#2î�è
ù �� ���½�#7�� 2î�è;&�¼�Ä\�P��IY�VPlß
�#M�ÁÂ��\���bc�Q��ï�àpd£i�µ�+ZL67á ¹
� 2î�è;&�¼�Ä\�� �P��IY uè[��½L6M�ÁÂ��W
���cß d£iLM���> IY¹U67��M¹��P�L:M�V�
�ÁÂ�� àpú9-lmÊ�c��¼�Ä �NüZ;�67
Ê�Ò%ôH��Q®��ï���j�gsL6Q}����ÁÂ���$
Á���<=ef�Q7���«�¢�®ð�&��$Á���5�Qµ \ß5
�67¼�Ä&�÷�����¸¢�¸ÑP�ºQ9Ì�8Ì �P��IY�V�
����«�¢�®ð�ß5�Q<DS ���ÁÂ��M��ë!±�¬î��
�b� ��$Á���5�Qµ . V#�ð��ñL6OlßL67���Á
Â��: lß&��B� ÁÂ���Ké�cL�"M���lßcL�" L
j\X67"Ì;lß�#M�&� Ì�bcL6lß`®�� eoO>?a/�
���������� ����������� ��
���
��6���ÁÂ��\¼�Ä lß�#cL679ÌM8Ì lßIY\�n�?;
X6M�&�lß&��å$Måú5��LH ÁÂ��\¼�Ä lß�#cL67
���ÁÂ�� �íN-IY ÷�)&��%IY÷¬Á���V� �%N�I
YL6�"M��%IY÷¬Á���V6ô&'ºr����½÷�);IYL6�
"��O�67�%IY÷�);&� ���ÁÂ��&���5�QË(mdMËm
�VP�¼�ÄO> lß-�;ÛÜ/ý)�QP�/��X6*�Q�Ú�Ë�
���QPL6:M\;�67���ÁÂ��&�½Ä&�%IY÷¬Á���+
� IYL6\��P��IYM��Q¼�Ä ÜXN-gcL6M���½÷�)�
,�L67
���ÁÂ��&�¼�Ä 9ÌM8Ì eo��� ),,( ZYX Ä�ÑPÒ%��
�Úáâ��� ),( YX �·´� �lß`®�� eo���iãM� ?��Öñ
-L67ñ-�Qeo���iãk\�=}Þ���M� .��Q��Q®�)Q
è®èW ÕÖ��ü�QM:<;���\R� �6O.#O b��#7!r�
7>�Q���iãO>� Z�M�®è�.�����iã\ôD�ýML6â
/Ð ���«!6V#�M6Ñ�#7: M��ÌÍ ���ÁÂ�� �µ/î
Meo Ö bc�±Wp�5�67á�����iã �0���D©���1�
ïo ÕÖ2�Ê�TML671�ïo\3o4U �Q>�� Q} �P��
IY\�±�QW Mb����P��IY Þ���©�L67���iãÊ�
TåM��A��ÅÆÑ� ÐÊ��wTÕÖ{ ��� ),( yx iã�cL675
¦2�u±¹ºM5¦F2�u±¹º��� lK SÊ5Z��cL67:��V
P��P��IY Þ���M� ��®è\�.5�QÊ�T 67�L�7
67á����ï�Ê8 D�p� �� D½67�\� D�/���ï�
ð D ),( yx �/��V#�d£i+ZL67
¹�&����V� gc5�Q�P��IY Þ��� uè[��M����
�?� 9�we� r�MÊ�§{�ÄN�ôUp�º6:M�V� äß6IY
;X67
!r����V� 7>�QÌÍÁÂ�� �µ/î :,¢;L6d£i�
� �ÌÍÁÂ�� �µ/îMeoM Ö bcT}67
á��¼�Ä �P��IYO>¹�Q®�)Qè®è<ÕÖ ª�ð lß`®�
�eo ��� ),( YX �cL67: ���M��ï����§¨��!=�D
T}67���*O>ûxÕ>Í�VP�?����) dct + �.L67á���
0���D�VPÊ� 'T �c����ï�Ê�T M §'T
T
m = �cL67��
�Ê� 'T �¹º�Q��ï� ��
)(),( mtBymtAx == �
M�����D©���1�ïoO>ÕÖ2� hgcL67:#� 7>�Q��
����
�
���������� ���������
���
)(),( nmtBynmtAx +=+=
¹�� IY �íuè[ML67�
û��� ),( yx @� ),( YX �¹ºL6Q} !=�
),(),,( tyQYtxPX == �
T}67
ð�¯HQÊ�Ò%ôH�������½W;�Q���ÁÂ�� Fm�u÷
_��V� |��ÌI;��lIOL�"��|��I����çuè[�V�
��lÂIL6�"M§¨�Q¿q����½W;� ��ÁÂ���lßTU
6�"&�ÌI V#�p��P;���ÁÂ��ô&�-� W�|��I���
�çuè[VPÓ�Fm\7>�6:M\±O�Q7
�
���������� ��������������
ÁÂ��\KÖMVP<Ú-®��ï���\;�6V#�L6Q}�&�K�°
_$Á¼0���� YÍ�/�÷������ KM_$Á¼0����;�6
V#�L6ÎÏ\X67í<=;&�_$Á¼0���� eW�^� ��GL6
¢ÁG��ýDX �ÁÂ��\A�/���p� �ðA �Ü���äº6ñ�
ðA�_$Á¼0�����Q�M�#ÜX�U6ñ�ðä�"#ñ-. �IM�
6V#�L6:M}ë�Q7�
ÁÂ�� _$Á¼0�����IäZL6Q} Ì\OPM-6Ãst��M
� &�bKB5�´?���ã ©\ÌÈÉ6���-. ��IY^ opÇ�6
:MM�Q7ÁÂ��\KM®��ï���L6M� Þ����C IY�´?
IY��/IY-. iãM� IY~�L6Q} NO\´?�bKB5�ã ©^�
"±é Þ�0�¬Ñ�Q®��ï������M� �<DK ®��ï���
���®��ï��� ��+� �HQ®��ï������»M� ��
�B��L67�
®��ï������ ²!PM� 6�5�Q�IËÀ���� Qî®��
±�¬î�� �Ip�L6Q}��®��ï������ ÏE oG�è
�����IDEM�@�º67G�è�O>7>�Q��M�IDE ��\®��
ï������ ���¢�nF5��: ®��ï������ GÌF\H
ÄN�Dg5�67�8 ®��ï������ ÆÇ\"HN;-�Mb`5�Q
M�&�é ���»: :,\�±�67ÆÇëmMb`5�Q�"&�ð´?\?
Iñ�ðã © �ºñ-.CÏ-Ãst_$Á¼0����ñê\�IDE¯H Q}
���¦���ñê5�67:#�Qì"N-b`�V� �G�è��\ê|�;�
�N-W ;X� W�ÁÂ��&�I¢�Q 6:M\;�67�
ÁÂ��&Kd5� �6ø>������� ®��ï������M�8
/�¡f� ù"L6:M�V� �AM ®��ï���\J�N-Þ���½
P;-��";W�®��ï������ZPKU ©ª�bc;�67��&O
PO�®��ï���������� á ¼�Ä �I=>� �6:M�W-
���������� ����������� ��
���
6 ;�¼�Ä �I=>� ®2ß �IL6:MW;�67�
~� /H§¨;&�ÁÂ���ú±� �6G�è;&A � IY�ÀL6
:M\;�-�7ª=�À�÷�����¸¢�¸-.�V� K IY�À�V#
M� W���ê���®è Q}�K IY �2 ?µ��O�À�O;�-�7
5>��K\L!M�-�IL6:M�V6ê�~F\äß67� V#-�";W�
ø>5�Qê��� ?µ�O>ÁÂ�� ÈIËØ�p�L6ÎÏ\X67:#�
Q/��bcL6Q}��IY�ÔN��L6:MM�Q7�Ô�I�&¦®¬�
�a��(��Q7�
ð�¯HQj���{�{��}Ç� ~°����F���Q7�
�
���������
����������� ����������������. �������� !"#
$. %&'(����)*. %&+,. -./.
Takashi TAJIMA, Yong XU, Toyoaki NISHIDA. Entrainment Based Human-Agent Interaction. To be
presented at 2004 IEEE Conference on Robotics, Automation and Mechatronics, December 1-3, 2004.
Traders Hotel, Singapore.
0�1�����23456789:����;<=>?@ABCDEF��G H 17I<JK
L)MNOPM� 1D1-05� 2003.
Q�RS�0�1�����2����;<;>?@ABCDEF�>TUVWX�H 17I<
JKL)MNOPM� 1D1-06� 2003.
����2<;����>YZ[��<=;����>YZ[��\]^_`a)>bcdef
g����hij_`kl� 44m�12n�pp. 1214-1220. 2003.
������� �!"#$%&'( �
� ËN-��gh���[\���lmL6Q}�&�gh¿q ´(N-3ß-.�
+¶ �����V63ßwQMU&�Nê{M¿"L6:M ÎÏF\�À5�W}
�67�O�����[EO> ��3ß j� <=�b� �¼�Ä\ÃstN
-Ì;Ç� ÜX�L6Q} i!N-<=&�AB �)'<= Åk;1U
>� �Q7�ý <=;&�AB �)'�§H &6O�IN-���ü\;�6
ÊÒ%�VP���üM�#´D3ß�Q:M;�st�V6_$Á¼0����
�|L6ÃstN-���ü ëmFß�Q7
���: jä; <=\�stN-W �V6_$Á¼0����Mü"5�6:
M;�¼�Älä� �����[ <=\VPjO-W �-6:M\��;�67
���������� ���������
���
)*+� ,-./�
� � � 01234� 56�789:;<=>�1234� ?@AB:;<=C�
�D����.EFGHIJKH� �
� ¢Á¬îï��â&Iâ;<=5�QHIMÏE��!"L6:M;�~��;
cÇ�OU6ÓFm;ÓMÙ-���[efL6:M�l�Q7oxÐz ¤ å
æ¼�hÈ�& §�¡£ ÓØÑ�_�Þ�lä ÍefM�#��¹º;-��oxoz�
oxyz <=Zq!"L6:M;~��QW ;X675>��íz !"~�;
&�\�^st¸H½��¢\ef�Q �hÈ� �\�^v2�Qî�½��¢\ef�Q�
Ö�wox|z{�#:M;�äåæ£�� P ��-²"�ÆÇL6:M;�
st¸H½��¢�v2�Qî�½��¢ <=!"L6:M;�89 ef� �
QÏE�� ��FßL:M�NM�Q7�
� åæ&�oxQz ì"���[++L6Q} êëì-��;XP���\~
Çð�� -�ÕÖ�;~�;�6:MßL:M&�í<=¢Á¬îï� ûCÏ�
�;X67�
åæ ¿q\���c MÙa��äð5é6:MwR�|�~��È��}z���Z��'���S�
��h�}z���Z��'���{��hÈ� �V6åæMÙ\��ûÓT_ W ;X6:M&L;�
±O� �6 ;�!"~�~�;&¸HÕÖ�ýDÌ Q7�
4J�~� õÏßL7�
L�MNOPQRN�
MEDLINEM&�NCBI (National Center for Biotechnology Information) \3nL6 P
���¦��;X67U£Vä/£��L6I �M2.P¡�� ^P�'���0
ÕD;�«êI D&õ [��� �O�üL67Þ�è b�&��®��^V×�¯�
�ï� �M�Q7 YZ�Z[�Þ�èb�M�Q ¢���¬��_�Þ� !©��ß
L7«êI �� #;��¯��ï�\(Çëm;X6W &õâD;X6:M\
±O67!Q�Ðp �¯��ï�XQP ÅÆ D&õ [� �Ð XQP ÅÆ
ÛtD&õ '�Ût;X67ìÛtD&õ [�²Ût�ü����st¸H;V�(Ç5
�6 Penn Treebank ÛtD\ [���t;X6:M1U6M��±} �67-_�
Þ�;X6M�U67�
I D 14,792,890
�¯��ï�\B8L6I D 7,434,879
ì D 70,815,480
ìÛtD 1,418,949,650
�� ������� �� � �����
��STUV W<XYZ[\]=�:�^_�
MEDLINE + åæ�&�Enju(ÇL67Enju& Penn Treebank WM�� +
+5�QÓWX- HPSG Í(Ç� ^P�5>�Y;åc5� �65!ë!-Ó
ØÑ��(ÇL6:M;�ÓMÙOpÓ�Ô-Þ�¬�½~�� �67
���������� ����������� ��
���
mu¬#� Enju& Penn TreebankM�#��ñû�®���Q_�Þ�(Ç�
Í£·\-5� ^P���� !!MEDLINEM�#ä/U£�����ÆÇ
L6M�tZ [��+ N- preference [��VPMÙ\���SJL67�:;�
í½��¢;&�GENIA _�Þ�(Ç� �£·÷�� Þ���ä/U£�
���Ç�ûÆÑL6wHara et al., 2005{7 3.7.2�ûÆÑI� GENIA_�Þ�ð
; + åæ MÙßL7Þ��� ûÆÑ�VPMÙ\���äð� �6:M
\±O67!Q�: ûÆÑ�^� &�Penn Treebank ;£·�Q÷���û�±
��Á1�Í�^º6ûI�6M� (ÇL6M�#ÌÍÇ� ^P�ÃÄ�x5-
©�_��;�Q-)�®� �����ûÆÑ;�6M�#Êk\X67
Penn Treebank GENIA Treebank
mu¬#� Enju 87.57 84.57
MEDLINIEÞ�èÇ Enju 87.31 86.02
�� ������� ���� ���������� !"#$�
�ú�����*���� W�MEDLINE Ç��Á�¼�½�QW (ÇL6
(Tsuruoka et al., 2005)7 3.7.3�£·Ç���M� <D _�Þ�(Ç�Q�"
¿qßL7GENIA _�Þ�M PennBioIE _�Þ�&.;>W¡®m�� _�Þ�
;X67�ý MEDLINE Þ�è�&ûJ; op £·_�Þ�LH (Ç�
£·��Q�*��Ç� �67
WSJ GENIA PennBioIE
WSJ 97.05 85.19 86.14
GENIA 78.57 98.49 86.59
PennBioIE 85.45 93.20 97.74
WSJ + GENIA 96.96 98.32 91.98
WSJ + PennBioIE 96.94 93.34 97.75
GENIA + PennBioIE 85.60 98.35 97.63
WSJ + GENIA + PennBioIE 96.89 98.20 97.68
�� ������� %&����'()*+�����,$�
HPSG Þ�Ä&���)m2��;Þ�è¿qcdL67���)m2��M&�
�������������í�M&é�ràL6��;X6\��L6����
& xmlMò�;X67Þ�ÄO>cd5�6���������&�.+GM¯t�%
+G;X67.+G��� &�xml;����ð�� !!�;�6Q}���&
-�\�¯t�%+G�p� &�.!Q\6ÄB5 ²B���L6ÎÏ\X
67�:;�¯t�%+G��� &���B� .�b� ?Ü id �T��±
P® id;ßL:M;�. ¨�4U6²B���� �67
`�abAB�
Enju\Ð + åæ�ÏL6ÕÖ&�AMD Opteron ¢ÁG�èw2.4GHz{ð;õ
ÐÇ;X6wNinomiya et al., 2005{7\]Få�ÓMÙ;�# HPSGÞ�ÄM� &
ÃÄ�ÓØ;X6\���;WMEDLINE��ÐS ¢ÁG�è;åæ�V#ML6
���������� ���������
���
M�õ 7000 �Ç�L-±;Ú0/� ÕÖ\OO6:M�-67�Q\� �: V
#-�67_�Þ�+ åæL6�"��� ¢ÁG�è(Ç� �ã�Þ�è
�#ÎÏ\X67
: Q}�Ç�Q©�µ¶&I¯�QíÐ�¸�Þ�MÑ�¸�Þ��;o5�Q
ï��("©õ 250��)��õ 170��)/340 CPU(Ç);X67�ã¸H& GXP
Ç� ��Q7�¯��ï�& 9,870K ^��A5�����è®è&"©;�
³Õ 3.2GB �_eÕõ 100GB;X67Õp ï��Ö�¥�2~®����[&-
�Q}�2~®�&dßN�_1�L6ÎÏ\X67Lï��Ö&Ý-�MW
100Mbps´Ù; ��� ö\ëm;X6Q}�LH 2~®�_1�� WD�
;XP�=>©�ÕÖM§H ÃÄ�é�ÕÖ�OOO>-�7�:;Û? Q}�L
H ���Lï�� 2~®�è�¡��)��ûI�_1��;o�Q7:�
�VP�Lï�� . ��);W�LH �¯��ï���ïG�;��¾¿
�¦\Û?�-67
~� �ã¸H^V×¾¿�¦&�GXP ���� �6Û?-��ï�«�Þ��
¸_»�)Ç� �±�Q7:�&�~��Q�_»�)LH �·Mp 2~®
���ñ¯� ^º&���>X� �6Ç���AP�� ~�L6M�#W ;�
�±q6¡���0¬Á�u�½���[M@Ò lm3nL67�� ��&P
GXP4] v2�Qî� ®������.�?ZÎÏMér�LAN!Q\�Q
µ¶�Ø�!"ëm;X67ÌÍ�Q¬�¯ ñê&2~®��-�Ò!�6Q}�
º�;©�l `Û�a��(�ï �`-.\R:� W�Û?��ß_»�) u
�� GXP �a� ~�5é6:M;�©� f�(ÌÍ� �-�_»�)) ~�
\ëm;X67
~�& 20050 9î 21® 1:21AM�eW5��� 9î 29®��LH ¸H\|Í
�Q7: ÖLH ¢ÁG�&ûS þ®Ù(niceo 19);~�5� ^P�� �O
¼�ÄW���[¥�(Ç� �Q7
�- �������./0123456789�:;<6'=45>?@�A�
���������� ����������� ��
���
X 3.7.1&~�eWO> ÕÖù¾MMW�ÌÍ�Q^ Dß� �67X 3.7.2&
íÐ�¸�Þ� ï��(8)�Ñ�¸�Þ� ï��(9) ���B� ï��
�� CPU�ÇÔ(ð)�Á�)�¦��¬(J�Æ?)ß� �67bc ×Ø&í~
�\�±� �Q�Öõ?dÖ;X67º�(9O>õ 1/3)?Ù�LH ¢ÁG�
eôÌÍ� ^P�� Ö(DÕÖ) Á�)\J\� �67!Q�íÐ�¸�Þ�;
M:<.:<Á�)�¦��¬� CPU �ÇÔ�1Pð\P\øù5�6\�:�&w
¼�Ä ¬�¯�V6W ;X67
�- �������B�C DEF'G�HIJK�L�
��������
(Yakushiji et al., 2005) Yakushiji, Akane, Yusuke Miyao, Yuka Tateisi and Jun'ichi Tsujii. (2005).
Biomedical Information Extraction with Predicate-Argument Structure Patterns. In the Proceedings
of the First International Symposium on Semantic Mining in Biomedicine. pp. 60—69
(Hara et al., 2005) Hara, Tadayoshi, Yusuke Miyao, and Jun'ichi Tsujii. (2005). Adapting a Probabilistic
Disambiguation Model of an HPSG Parser to a New Domain. In the Proceedings of IJCNLP 2005. (to
appear)
(Tsuruoka et al., 2005) Tsuruoka, Yoshimasa, Yuka Tateishi, Jin-Dong Kim, Tomoko Ohta, John
McNaught, Sophia Ananiadou and Jun'ichi Tsujii. (2005). Developing a Robust Part-of-Speech Tagger
for Biomedical Text. In the Proceedings of the 10th Panhellenic Conference on Informatics. (to appear)G
������� �!"#$%&'( �
� fÖ��M�#�st£HI G�O> Í�~ �����ÆÇ��ø�Ü÷
+G©�L6¸H\Ðy²tM�#���-����²"�b� ÆÇëm;X6:
Mß�Q���;ûJ W ;X67:�!;&��67M�±�6<=;W�
¡����{����~}�M4&�6g�¸HDè�t �����ÆÇL6´Ù;X�Q7�
ý !"~�&�+¶ ��Ë�� L�;ª�N-W ;X67�
� !Q�: ~�&�: V#-�67~�\�©�l§¨M� &Êé-_úL6
:M-��®ÄN-�¥M� ~�;�6:Mß�QDWCÏ;X67~� cÇ�
�;&�¢}���h}� LH ����¸HL6:M&�?Ù~�L�&V��XM
&��Q-À�wQMU&�Ðíî�{ ¸H�#¹º;Ï�;XP��8 §¨
���������� ���������
���
�U&�DÕÖ ®ÄN-�¥M-67: D;&�89 <=Zq&�L;��~
c��;�U6W M-� �67�
� 89 �m�L-±;�ð����²"�b� X>Oß}?eN-st¸Hh��
��¬L6:M;�VPÓÙ-ËÀ¸HM¿"L6H�;X6ñM�#�m&�:
åæ¿qI3��oxÚz hÊ+G��QËN-gh�X6�&���ïå
M ���c��efL6:M;üZ5�67: V#-ì"���[ ++p�
&�áz;�þL6\�: !"~� Zü&�89 �ýN-ÌJ <=+,\�~
N�~�ëm;X6:Mß�Q:M�-67�
� �8�¢}���h} �� åæ¿q�����M� �eL6:M=�� ^P�
oxyz £��¬ã_�Þ�MMW��äåæ£�� ����»®¼�½ <=��
N��=L6�����M� ÐÇ5�6:M��� �67�
�
)*c� ,-d<ef�fg^�Zh�
� � � � 0��:;<=ijC�
�D����.EFGkK� �
� í¢Á¬îï� jÍIN-Êà&�st¸H�m��Á¬����±�¬î���
��v2�Q��M�#��� �-6��!"L6:M;������ ��
¼�Ä�M� ��-��¼�Ä�M� ���L��;3ßL6è�i���
�[++L6:M;X67�
� : V#-!"è�i����[ ++&�ÏE��¼�Ä���[��%ðÉ6
!"��\ÎÏM-6¹º;-��¼�ÄM jk-�d�±¿×�O�> ÏT
Ï���æL6ÎÏ\X67�
� í¢Á¬îï�;&�¢Á¬îï�=� ÕD;ëm-¼�Ä���[++��
Æl�eL6M�Õ��~¼�ÄM� äåæ£�� <=>²³wÊ�<=ðm�[
è®±��ñ ½��¢�÷���<=Ú�n('Õ��-.{M ¥�<==}
�Q7: V#-<= ZqM� �\�^��Çt�cè�i����[�\�^�stÇ
Wgh���[�\]^opq�r�#�± �I�c���[ef��?e��e�
�674J;&�:�> �e���[�p� ¯H67�
L�lm�@�nopq�fg^r?s�
m��Á¬�½��¢\f_5é �Q��Çt �I�À��çuè[�Web
O>(Çëm-���[^V×"Q�Á�)ëm-÷¬Á��M� ½ú�Q7!Q�
�ßÌÍ®ít4]�¤t�� w s�Á�Þ tst�� t�l��Q7:
���[ðs_ WebñM�#-u;�e�Q7�8�Çt�c 3000 �ïG�/î�
÷¬Á��"®�Á�) 200O/î;X67X 3.8.1�s_Web ®��2î�� ª�
ßL7
�� ��Çt�c <= ��&���_�Þ�l¶£·L6jÍ\vÇ5�6
:M\��\�s_ ª}� ��çuè[&_�Þ�²BF�st²BFëm-�P�
���������� ����������� ��
���
���,-uv��W�±r��"�V� &�îEåæ���[5U��L6:M\
ëm;X6M�#áw5ðML67V� �ðñ V#��st: ÆÇ&éÕÖ
;ëm�-�Q7!Q�: s_���[&�m��Á¬�½��¢ <=Zqxá
!"� ��¢��2y�[M� �±� ^P��8�Û?-ÌÍ ��Çt�c
O>®ít� P�P#t ÑPÒ��#<=Zq-.\ÑPÒ!�� �67�
s
s�- ��M����NO P#Q �RSTUVWX�
�
��t?@�uvw�fg^�
WWWð ����b���Qst²BF -� Üã¦�� ÇWgh���
[Kiwi++����zÌ�Ë�c{���[M�M ëmF���Q7Kiwi gh
ª�X3.8.2�ßL7
�
�-��M����YZ[Z�\]WX^_`.ab�
�cdefgh#�i#j�Z��kZlhg#"�ghmjf��nopqrNT�s�\]$�
s
���������� ���������
���
�ß|}ºÁ�P�-����2~®��ÆÇ��ËÀ»®¼�½ Q} ~�
2Ò%ß�#Portable Kiwi$Y�Q7WU&�ð�V6ñM�#� I��
��6 ÜãKiwi ÌÍ;²õãÈL6:M�V� û�M¿q bO>-6ËÀ
~�»®¼�½;�6V#�-�Q7
`�xyz{|}~���\]�fg^�
s ����»®¼�½ �A ?p&������ ��½H�Ïõ5�Q�;¼�
Ä�3ßL6:M;X67U£ä/£���^� &�WU&�r�#���ÞïË�
&'Ý�-. ���±\�c b�M-P#6CÏ-ËÀ;XP���&a��(�
ï ��;�L6:M\;�6wX YZ�ZY{7��)&�r�#�Q2��ËM��
Q±��«�«�bc��±�¬\��> Ö �±�bcL67±�¬��T5�6
��&���+ZL6±��«�« ï��V� �-P��"�V� &��±
�³¹º;&-���± �@�positive/negative�confidence -.��L6��\p
!�67�
�
���
�����
��
�
�
�- ��M���� tuvwx�yz{|�}�
�
� 89&L;�wYakushiji et al., 2005{�^� �¯t�%+G(ÇL6:M;�Ó
�MÙ;��ÞïËÖ��YÇ�c;�6:M��� �67!Q�pqMr�#
Ö �±�p� W�st¸H��Ml¶£·Ý%"±é6:M;�̈ �Ô��
�.uML:M-��Ó�MÙ �±�c\;�6:M��� �6wChun et al.,
2006{7
� : V#-<=Zq�~� ¡®m<=>\ð�U6ñV#�L6Q}�&��c
5�Q���½H�!"� ¼�Ä�3ß��5>��¼�Ä ´D�ËÀ#��
Q�;7>�QËÀ½H�-^L:M\;�6���[++L6ÎÏ\X67�
¡®m�� Q} AB ����»®¼�½���[M89 ���[M [�&�
���������� ����������� ��
���
ø�+ åæ ¿q(ÇL6:M;�+¶ÌÍVPWMÙ Ó��±�c\ëm�-
� �6:M�WX6\�GCL ^V×EF+G���¦��(ÇL6:M;���
� Ü÷+Gl��Qgh lm¼�Ä�3n;�6D�WX67�
�� ��
� !"����»®¼�½���[ ~�jÍM� &�ä/£���¦�� ®��
�2î��-.�°�(Ç5� �6�Web �¢u0����M�#�î(ÇL67
Web�¢u0����&�̄ QÄ5UÇÜ;��&.:;W�#:M\;��LH
lm\è�¡E;3n5� �6Q}�lm LM�¹�W�Ø;X67X YZ�Z��
í!"���[ +ZßL7�
� Webè�¡;&�JSPM Ser��}z�V� ¼�ÄM ®���2î��~�� �
67¼�Ä& ª}�¯QĽ� ��� gh�rB-.�LH ÈY�#:M
\;�67�
�
Web����
(JSP, Servlet)
DB����
(SQL)
GCL��������(Web����)
�� ��������
�
�- ��M�~�� ���������
�
�������
GCLè�¡�&�xml;�5� �6MEDLINE �����\õ�5� �67
�^º6.+G�¯t�%+G�����L6��\®���ï�5� ^P�
GCL ï±u��V� ¼�Ä\l��Q����ÓØ�ghL6:M\;�67Q
MU&�GCL(ÇL6:M;ðI� �]z��} Ü÷ð �tM¯t\Q2��Ë;X
6 ñ-.ghL6V#-:MWëm;X6wX YZ�Z�{7: V#-gh&Ût
®���ï�¦��ML6½Ä gh±�¬�;&~�L6:M&y,;X67�
GCL è�¡\�����bL6���������rà� �6 �b���� è
�¡;&�����M&PG�Q��ràL6�Aàp7��N�&�±��«�
«��L6!©���¼�Ä\++�QËÀa��(�ï-.ràL67QMU&�
L;�����O>�c5� �6pqMr�# �±��wX YZ�Z�{-.&�DB
è�¡�;rà5�67�
���������� ���������
���
�
�- ��M����}��f�)�m gZ�mg#fpU�r���
�
�
�- ��M����������������
�
��yz{|}~���g�f�������
s pqMr�# �±�cL6:M&�U£�ä/£���^º6����»®¼�
½ CÏ-�� ?p;XP�� V#-�±����O>fnL6Q} ���[
\�ArrowSmith, BITOLA-.&ß}M� L;���pO~ÇÑ5� �67AB
�±fn���[ ��&��±�� �cÇt�° ¥R����� �6Q}
��7>�Q¿q��� �®è\p!� �6M�#��\X67BIONLP ��;
&�l¶£·M���¦�� ÌÍ-.Ý%"±é �MÙ Ó����c�#Ì
ÍW<=5� �6\��c¿q\���� �;6ß �6:M\���� �"�
�c¿qMä/£���¦��M\u�ï� �-�7�c¿q\�U£ä/£<=
���������� ����������� ��
���
ð;�ÇOp¨(Çëm;X6Q}�&������;�À5� �6±��uM�ä
/£���¦�� ±��uM Ö�bc\p� �6:M\êëì;X67�
� í<=;&�ä/£���¦��O>++�Q,-Ml¶£·ÌÍÝ%"±é6:
M�VP�ä/£���¦��Mbc\p� �6r�#�pq �±���ÓMÙ
;����O>�cL6:M�NML67�
� X YZ�Z���í�¢Á�� ��=ßL7�
�
Disease DictionaryGene Dictionary
MedLineMedLine
HUGO DDBJLocusLink SwissProt RefSeq UMLSHUGO DDBJLocusLink SwissProt RefSeq UMLS
GeneDiseaseRelation GeneDiseaseRelation
Filtering
Disease-Gene Pairs
�
�- ��M���� ������������
�����
s r�#,- ++�&�|p Þ¯u�ï���¦��wHUGO, LocusLink, SwissProt,
RefSeq, DDBJ{(Ç�Q7±��u D&�34,959±��u;X67
� pq-,-�&�UMLSw2003AC Edition{(�Q7UMLS MetathesaurusO>�
Semantic�®¢M� �pq��/��L6 TUIWp±��u�c�Q7� ¿q�
Üã DM� �Y[��'�K�õöDM� [������O>-6pq-,-\++5�Q7�
�������� !
� l¶£· ���¼�½M�� Q}����������_�Þ�YZ�Q7û
J�MEDLINEO>á ï±u��V� �pq��±L6�¯��ï��cL67
“Disease Category”[MeSH] AND (“Amino Acides, Peptides, and Proteins”[MeSH] OR
“Genetic Structures”[MeSH])
� 7>�Q [�Y�'�'���¯��ï��b� �r�#�pq,-M�ûk?IÍ(
ÇL6:M;������ r�#-~��o-~� �À���r�#Mpq ¥
���������� ���������
���
R ~�²"Y67: �O> [��� K ¥R�"[è�¢u�½� ����
�����#b�M�Q7�����&�r�#-~��pq-~��b� ���
>\�M��r�#-wX6�&o-{M� �À5� �6O.#OÚo;���
�����#75>��r�#-�pq-\MW�M���"�� ¥R\r�#�
pq�±� �6O.#O�p� WÚo;��������#7r�#\pq
��±� �6O.#O b`�p� &�4J o�@ �± �r�O\ñ¯5�
�6O.#O�_M�Q7�
i�oîäH£�pq �P¼è[�
i�r�#W��&��ÞïË ��N�q�
i�pqu�ï��`�=��`�^º6»�P�M� �A�
� ������� ¿q�r�#-Mo-\LjMWM���À5� �6�"����
A";�� ¥R\r�#Mpq �±�� �6:M\d>O�-�Q7�
Q}�í��ï�^� &�2«��u�½ b�&�r�#�pq- �À %M��
r�#Mpq �±��� &2«��u�½&�±-�7�
"#$%&'()*+,-!.
� Çt�À 2«��u�½ Q} l¶£·ÌÍ�&�û�±��Á1�Í(Ç�
Q7bé Q} Êà�M� &�4J ��Ç�Q7�
i�r�#-�À�
i�r�#-~���Ê8 unigram/bigram����head�~� argumentM� �
6¯t�expanded formw~�\ acronym �"{
- pq-�- pq-~���Ê8 unigram/bigram����head
/012
� YZ�Z[�pq-�r�#- �ÀMÙ [�i�{�� �9g�;���Q¿qßL7
::;�¨�Ô;&-��b¨�Ô��l�M� �6 &��W�W������
��_�Þ�Y6���,-»��;~� �À�� �6Q}��� Ü÷;
¨�Ô&©�;�-�O>;X67¿qn6M�r�#-�pq-MW��Ó�MÙ
;l¶£·�V6b�\;� �6:M\±O67�
�
� Æ"Ô� �b¨�Ô�
r�#-� ��Z�� ��Z��
pq-� ��Z�� ��Z��
�� ��M���� ���������x���
�
� á��r�#�pq�± �c��L6MÙ� ����� �ßL7®�¯HQV#��
�c���[&�r�#-�pq- .;>MW\M��Mb�5�Q�"�³NO�
��> Ö��±\X6M%-L7¿qn6M�Æ"Ô\ ��y´ÙO> ��y/�:
M���äð� �6:M\±O67�b¨�Ô&ÐA´Ù�OSJ� -�7�
���������� ����������� ��
���
� Æ"Ô� �b¨�Ô�
2«��u�½-�� �[Z��� [��Z���
2«��u�½XP� ��Z��� ��Z[��
�� ��M���� ��������������
��������
(Chun et al., 2006) Chun, Hong-Woo, Yoshimasa Tsuruoka, Jing-Dong Kim, Rie Shiba, Naoki Nagata,
Teruyoshi Hishiki, and Jun'ichi Tsujii. (2006). Extraction of Gene-Disease Relations from Medline
using Domain Dictionaries and Machine Learning. In the Proceedings of the Pacific Symposium on
Biocomputing (PSB) 2006. (to appear)G
������� �!"#$%&'( �
��!"è�i����[&�ef5�QÏE��\~� ¼�Ä ÏT��5é6�
��[++���;X6:M��L6M�Õ��á�<=LH��� jän�}
6ð;��-�AqQL7�
íz opq�r�#�± fn£��#���[&�÷�ì"<=Ú�n('Õ
� ���½��¢M ¥�;��QW ;XP�U��¡®m<=>\�ÇL6:M
;�¼�Ä ÏTÑPK�QW �� ��=�;X67M������»®¼�½
M����¸H�¡®mm��Á¬�M VP�lN-¿"� D1V�Q�± �
À-.��Q-<=��»\��5� �67: jä; <=&�äåæ£ Q}
����»®�½ <=M� ����°\PppX67�
!Q��ý !"è�i����[;&�¼�Ľ��¢M <=�d\W�MW=
2;�6opq�r�# ��ÑPðÉQ\�����O> ��Ë��YÇ �I
�c�£��� m��Á¬��V6��Ë lm�� Úp���[\��ß �hÈ� å
æ¿qWM�ef5� �67�
:�> �-�Q��ï Q}}è�i����[\��ßst¸H ¿q���
+Z;�6:M&���ÏTÕD4I��O-P µ� st¸H�67�����
ÆÇ���ËN-hÊ+G�#�5é6:M;��ÔN�Op�ÓMÙ�ËN-¸H
è�i�ÕD;ëm�L6M�#�í¢Á¬îï� ÌJ �ýN-��\�~N-
��;X�Q:M� �67�
���������� ���������
���
����������
�D��?@AB:;<=�
oG pG qG rG sG tG uvwxG yz{|G
}~��G��P)P)�G
_`)�G��G ��kl;��G
���������
��������G
����G��P)P)�G
_`)uv)aG��G
����D����
hiG
���������
��������G
����G��P)P)�G
_`)�G��G � ��D ¡¢£G
���������
��������G
¤¥¦§�G¨©ª«hiG
P)�P)G���G ¬D®D$>¯°G
���������
��������G
±²³´G��P)P)�G
µ¶·¸�a)G��G ¬D®D$>¯°G
���¹�Gº��
��������G
»¼½¾G��P)P)�G
_`lJ)¿uvaGÀÁÂÃÄuvÅ Æ����D>6�G
���¹�Gº��
��������G
ÇÈÉÊG��P)P)�G
_`lJ)¿uvaGÀÁÂÃÄuvÅ ËÌ)ÍG
���º�Gº��
��������G
Î�ÏG��P)P)�G
_`lJ)¿uvaGÀÁÂÃÄuvÅ _`ÐÑG
���¹�Gº��
��������G
ÒÓÔGÕÖÔ×GØÓÙG��P)P)�G
_`lJ)¿uvaGÀÁÂÃÄuvÅ _`ÐÑG
���¹�G���
���Ú�G¹�G
����G��
��������G
P�ÜÝG��P)P)�G
_`lJ)¿uvaGÀÁÂÃÄuvÅ ?DÞ$>6�G
���Ú�Gº��
��������G
ß=à�G¨©ª«hiG
P)�P)G��G _`ÐÑG
���������
��������G
áâã´G¨©ª«hiG
P)�P)G)äuvÅ � ��D ¡¢£G
���������
��������G
åæçeèéG��P)P)�G
_`lJ)¿uvaG�êëìG _`ÐÑG
���������
��������G
íîïðG��P)P)�G
_`lJ)¿uvaG�êëìG � )ÍG
���������
��������G
ñòàÝG��P)P)�G
_`)�G�êëìG ó5ôõG
���Û�Gº��
��������G
È�ãöG��P)P)�G
_`lJ)¿uvaG�êëìG _`ÐÑG
����G���
��������G
÷�øùG��P)P)�G
_`lJ)¿uvaG�êëìG
!ú$�¬D®D
$G
���¹�����
��������G
û�üùG��P)P)�G
_`lJ)¿uvaG�êëìG ¯�ýþG
���º�Gº��
��������G
N��G��P)P)�G
_`lJ)¿uvaG�êëìG _`ÐÑG
���Û�Gº��
��������G
���������� ����������� ��
���
��G��P)P)�G
_`lJ)¿uvaG�êëìG Æ����D>6�G
�����Gº��
��������G
����G��P)P)�G
_`lJ)¿uvaG�êëìG ËÌG
���º�����
��������G
����G��P)P)�G
_`lJ)¿uvaG êëìG Æ����D>)ÍG
���º�����
���Û�G¹�G
��G��P)P)�G
_`lJ)¿uvaG�êëìG ËÌ)ÍG
���º�����
��������G
P���§G��P)P)�G
_`lJ)¿uvaG�êëìG � ��D ¡¢£G
���º�����
��������G
��1G��P)P)�G
µ¶·¸�a)uva�êëìG �Â�E$!£G
���º�GÚ��
��������G
���G��P)P)�G
_`lJ)¿uvaG êëìG � )ÍG
���¹�����
��������G
��)G��P)P)�G
_`lJ)¿uvaG êëìG � ��D ¡¢£G
��������
��������G
���ÝG��P)P)�G
_`lJ)¿uvaG êëìG � ��D ¡¢£G
��������
��������G
��u�G��P)P)�G
_`lJ)¿uvaG êëìG Æ����D)ÍG
��������
��������G
û��� G��P)P)�G
_`lJ)¿uvaG êëìG _`ÐÑG
���Û�Gº��
��������G
µ~!8G��P)P)�G
_`)uvaG êëìG
���>�D���
�hiG
����G��
��������G
P"#�G��P)P)�G
_`)uvaG êëìG
���>�D���
�hiG
����G��
��������G
$%�G��P)P)�G
_`)uvaG êëìG
���>�D���
�hiG
����G��
��������G
&'(G��P)P)�G
_`)uvaG êëìG
���>�D���
�hiG
����G��
��������G
)�*G��P)P)�G
_`)uvaG êëìG
���>�D���
�hiG
����G��
��������G
&+P,G��P)P)�G
_`lJ)¿uvaG êëìG � ��D ¡¢£G
��������
��������G
-�à�G��P)P)�G
_`lJ)¿uvaG êëìG _`./G
��������
��������G
Èî�P0G��P)P)�G
_`lJ)¿uvaG êëìG ËÌ)ÍG
��������
��������G
12¾� G��P)P)�G
_`lJ)¿uvaG êëìG _`ÐÑG
�����Gº��
��������G
�%34G��P)P)�G
_`lJ)¿uvaG êëìG _`ÐÑG
�����Gº��
��������G
56Ó7Ó88G
Ã89Ô×:;G
��P)P)�G
_`lJ)¿uvaGuv<G � ��D ¡¢£G
�����Gº��
��������G
���������� ���������
���
=×>?:Ô@>>G
Ä6>?G=×9ÔG
��P)P)�G
_`lJ)¿uvaGuv<G ËÌG
�����Gº��
��������G
A�BG��P)P)�G
_`lJ)¿uvaGuv<G ?DÞ$>6�G
����G��
����GC�G
âîD��G��P)P)�G
_`lJ)¿uvaG êëìG ó5ôõG
���¹�Gº��
���Û�G¹�G
E���G��P)P)�G
_`lJ)¿uvaG êëìG � ��D ¡¢£G
���¹�Gº��
���Û�G¹�G
F�GHG��P)P)�G
_`lJ)¿uvaG êëìG ËÌ)ÍG
��������
���Û�G¹�G
"î�ÝG��P)P)�G
_`lJ)¿uvaG êëìG
<Ia)>G
!ú$�®D$6�G
���º�GÚ��
����G�G
��JK½G��P)P)�G
_`lJ)¿uvaG êëìG
<Ia)>G
!ú$�®D$6�G
���º�GÚ��
����G�G
LòM±G��P)P)�G
_`lJ)¿uvaG êëìG óN��klO��G
���º�GÛ��
����G�G
P�(�G��P)P)�G
_`lJ)¿uvaG êëìG
<Ia)>G
!ú$�®D$6�G
���º�Gº��
����G�G
s
������56�789:;<=s
oG pG qG rG sG tG uvwxG yz{|G
Q¥�*G��P)P)�G
_`lJ)¿uvaG��G R·STU5O��G
���������
��������G
��¦V G��P)P)�G
_`lJ)¿uvaG���G R·STE$!£G
���������
��������G
÷� WG��P)P)�G
bX�YuvaG���G U5Z[G
���������
��������G
"\]G��P)P)�G
_`lJ)¿uvaGÀÁÂÃÄuvÅ
R·STE$!£>
¯°G
���¹�G���
��������G
�^_`G��P)P)�G
_`lJ)¿uvaG êëìG
R·STE$!£>
¯°G
��������
�����G¹�G
a�bcG )äG \duvÅ R·STE$!£G���������
���º�G¹�G
eàfG��P)P)�G
_`lJ)¿uvaG êëìG �Â�E$!£G
���������
���º�G¹�G
g%h<G��P)P)�G
_`lJ)¿uvaG êëìG �Â�E$!£G
���¹�����
���Ú�G¹�G
a�i�G��P)P)�G
_`lJ)¿uvaG êëìG �Â�E$!£G
�����Gº��
��������G
áîjkG��P)P)�G
_`lJ)¿uvaG êëìG �Â�E$!£G
�����Gº��
��������G
���MG��P)P)�G
_`lJ)¿uvaG êëìG
R·STE$!£>
¯°G
�����Gº��
��������G
���������� ����������� ��
���
lmnG��P)P)�G
µ¶·¸�a)uva êëìG
R·STE$!£>
¯°G
�����Gº��
��������G
s
���������<:;<=s
oG pG qG rG sG tG uvwxG yz{|G
��MoG��P)G
_`8pq�DG��G
!ú$�èr>Æ�
���Dó5stG
���Ú�Gº��
��������G
��D�G u�vOwP)G ��G����DxlE
$!£G
���������
��������G
��S±ÝG��P)G
_`lJ)¿uvaG���G
yz./E$!£>
{ZG
���Ú�Gº��
��������G
á�|G��P)G
_`8pq�DG��G
�Â�}D�¯~ýþ
)Í���¡¢£G
���������
��������G
����G��P)G
_`8pq�DG��G
yz./E$!£>
��G
�����Gº��
��������G
�î�ÝG��P)P)�G
_`)�G êëìG �Â�èr>_`ÐÑG
����GC��
��������G
�~ÏG��P)P)�G
_`)�G êëìG
yz./E$!£>
{ZG
����GC��
��������G
�^k�G��P)P)�G
_`)�G êëìG
yz./E$!£>
�yG
����GC��
��������G
&�G��P)P)�G
_`lJ)¿uvaG êëìG
<�Æ����DÐ
Ñ>.�G
����GC��
��������G
��òRÝG��P)P)�G
_`lJ)¿uvaG êëìG
yz./E$!£>
�yG
�����Gº��
��������G
���G��P)P)�G
_`)�G êëìG
yz./E$!£>
�yG
�����Gº��
��������G
���G��P)P)�G
_`)�G���tÅ
yz./E$!£>
��G
����G��
����G��G
Ø79>�G�Ö��G��P)G
_`lJ)¿uvaGÀÁÂÃÄuvÅ
����D���
E$!£>6�G
���¹�Gº��
���Ú�G¹�G
��P�G��P)G
_`lJ)¿uvaG êëìG Æ����DJ)G
���º�Gº��
���Ú�G¹�G
%���G��P)P)�G
_`)�G êëìG ��y�ÐÑG
����G���
�������G
�^�<G��P)P)�G
_`)�G êëìG
yz./E$!£>
��G
����G��
���Û�G¹�G
} ¡G��P)P)�G
_`)�G êëìG ��y�ÐÑG
����G��
���Û�G¹�G
�ÏG��P)P)�G
_`)�G êëìG
yz./E$!£>
��G
����G��
�����G¹�G
¢�£G��P)P)�G
_`)�G êëìG ��y�ÐÑG
���Û�Gº��
�����G¹�G
���������� ���������
���
�� �� �
�*���[����Y�F� D� �H��� �� �_�
1. Kenjiro Taura, Andrew Chien: A Heuristic Algorithm for Mapping Communicating Tasks on
Heterogeneous Resources. In Proceedings of Heterogeneous Computing Workshop 2000,
pages102-115, 2000.5
2. Takashi Ninomiya, Kentaro Torisawa and Jun'ichi Tsujii. An Agent-based Parallel HPSG Parser
for Shared-memory Parallel Machines. Journal of Natural Language Processing. 8(1). 2001.
3. Hideki Mima, Sophia Ananiadou and Goran Nenadic. The ATRACT Workbench: An Automatic
Term Recognition and Clustering of Terms. In the Text Speech and Dialogue (TSD 2001), Lecture
Notes in Artificial Intelligence. 2166. pp. 126--133. Springer Verlag. 2001.
4. Kenji Kaneda, Kenjiro Taura, Akinori Yonezawa: Virtual Private Grid: A Command Shell for
Utilizing Hundreds of Machines Efficiently. In Proceedings of 2nd IEEE/ACM International
Symposium on Cluster Computing and the Grid (CCGrid 2002), pages212-219, 2002
5. Naoki Yoshinaga and Yusuke Miyao. Grammar conversion from LTAG to HPSG. WEB-SLS:
the European Student Journal on Language and Speech. 2002.
6. Jin-Dong Kim, Tomoko Ohta, Yuka Teteisi and Jun'ichi Tsujii. GENIA corpus - a semantically
annotated corpus for bio-textmining. Bioinformatics. 19(suppl. 1). pp. i180-i182. Oxford
University Press. 2003.
7. Naoki Yoshinaga, Yusuke Miyao, Kentaro Torisawa, Jun'ichi Tsujii. Parsing Comparison across
Grammar Formalisms Using Strongly Equivalent Grammars. Journal of Traitement Automatique
des Langues. 44(3). pp. 15--39. Association pour le Traitement Automatique des LAngues.
2003.
8. Toru Hisamitsu and Tsujii, Jun-ichi. Measuring Term Representativeness. In Pazienza, Maria
Teresa (Eds.), Information Extraction in the Web Era. LNAI 2700. pp. 45-76. Springer-Verlag.
2003.
9. Minoru Yoshida, Kentaro Torisawa, Jun'ichi Tsujii. Extracting Attributes and Their Values from
Web Pages. In A. Antonacopoulos, Jianying Hu (Eds.), Web Document Analysis: Challenges and
Opportunities. pp. 179-200. World Scientific. 2003.
10. Hiroshi Nakagawa, Tatsunori Mori. Automatic Term Recognition based on Statistics of Compound
Nouns and their Components, Terminology, Vol.9 No.2, pp. 201-219, 2003/4
11. Kenji Kaneda, Kenjiro Taura, Akinori Yonezawa: Virtual private grid: a command shell for utilizing
hundreds of machines efficiently. Future Generation Computer Systems 19(4), pages 563-573,
SanDiego, 2003.5
12. Kenjiro Taura, Toshio Endo, Kenji Kaneda, Akinori Yonezawa: Phoenix : a Parallel Programming
Model for Accommodating Dynamically Joining/Leaving Resources. In Proceedings of ACM
SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPoPP 2003), pages
216-229, 2003.6
13. Takashi Hoshino, Kenjiro Taura, Takashi Chikayama: An Adaptive File Distribution Algorithm for
Wide Area Network. Proceedings of Workshop on Adaptive Grid Middleware,New Orleans,2003.9
14. Takashi Hoshino, Kenjiro Taura, Takashi Chikayama: An Adaptive File Distribution Algorithm for
Wide Area Network. Spcial Issue of Journal of Parallel and Distributed Computing Practices,To
appear ,2003
15. Minoru Yoshida, Kentaro Torisawa and Jun'ichi Tsujii,G Extracting Attributes and Their Values from
Web Pages,Chapter in a book titled Web Document Analysis: Challenges and Opportunities, A.
Antonacopoulos and Jianying Hu, editors, Series in Machine Perception and Artificial Intelligence,
World Scientific, pages 179-200, 2003/9
���������� ����������� ��
���
16. Yusuke Miyao, Takashi Ninomiya and Jun'ichi Tsujii. Corpus-oriented Grammar Development for
Acquiring a Head-driven Phrase Structure Grammar from the Penn Treebank. In Keh-Yih Su,
Jun'ichi Tsujii, Jong-Hyeok Lee and Oi Yee Kwong (Eds.), Natural Language Processing - IJCNLP
2004. LNAI3248. pp. 684-693. Springer-Verlag.
17. Yoshimasa Tsuruoka and Jun'ichi Tsujii. Improving the Performance of Dictionary-based
Approaches in Protein Name Recognition. Journal of Biomedical Informatics. 37(6). pp.
461-470. Elsevier. 2004.
18. Yoshida, Minoru, Kentaro Torisawa, Jun'ichi Tsujii. Integrating Tables on the World Wide Web.
Transactions of the Japanese Society for Artificial Intelligence. 19(6). pp. 548-560. The Japanese
Society for Artificial Intelligence. 2004.
19. Takashi Masuyama, Hiroshi Nakagawa: "Two Step POS Selection for SVM based Text
Categorization", IEICE Transaction of Special Issue on Information Processing Technology for Web
Utilization, Vol. E87-D, pp.15-21, 2004/2
20. Toshio Endo, Kenji Kaneda, Kenjiro Taura, Akinori Yonezawa: High Performance LU Factorization
for Non-Dedicated Clusters,IEEE/ACM Symposium on Cluster Computing and the Grid
(CCGrid2004), Chicago, pp.678-685, 2004.4
21. Kenji Kaneda, Toshio Endo, Kenjiro Taura, Akinori Yonezawa: Routing and Resource Discovery in
Phoenix Grid-Enabled Message Passing Library, IEEE/ACM Symposium on Cluster Computing and
the Grid (CCGrid2004),Chicago,pp.670-677,2004.4
22. Kenjiro Taura: GXP : An interactive shell for the grid environment. In proceedings of Innovative
Architecture for Future Generation HighPerformance Processors and Systems
(IWIA2004)G ,Hawaii,pages59-67,2004
23. Minoru Yoshida and Hiroshi Nakagawa: "Specification Retrieval - How to Find Attribute-Value
Information on the Web", Natural Language Processing -IJCNLP 2004, Keh-Yih Su Jun’ichi Tsujii
Jong-Hyeok Lee OiYee Kwong (Eds.), Lecture Notes in Computer Science, Springer Berlin
Heidelberg Vol.3248,pp.338-347, 2005/10
24. Toshio Endo, Kenjiro Taura: Highly Latency Tolerant Gaussian Elimination, Proceedings of the 6th
IEEE/ACM International Workshop on Grid Computing¤Grid2005¥¦Seattle¦pp. 91-98¦2005.11
25. Hideo Saito, Kenjiro Taura, Takashi Chikayama: Collective Operations for Wide-Area Message
Passing Systems Using Adaptive Spanning Trees, Proceedings of the 6th IEEE/ACM International
Workshop on Grid Computing ¤Grid2005¥¦Seattle¦pp. 40-48¦2005.11
26. Yuuki Horita, Kenjiro Taura, Takashi Chikayama: A Scalable and Efficient Self-Organizing Failure
Detector for Grid Applications, Proceedings of the 6th IEEE/ACM International Workshop on Grid
Computing ¤Grid2005¥¦Seattle¦pp. 202-210 ¦2005.11
27. Irena Spasic, Sophia Ananiadou, and Jun'ichi Tsujii. MaSTerClass: a case-based reasoning system
for the classification of biomedical terms. Bioinformatics. 21(11). pp. 2749-2758. Oxford
University Press. 2005.
28. Jun-chi Tsujii and Sophia Ananiadou. Thesaurus or logical onotology, which do we need for
mining text? Language Resources and Evaluation. 39(1). pp. 77-90. Springer SBM. 2005.
29. Jun'ichi Kazama and Jun'ichi Tsujii. Maximum Entropy Models with Inequality Constraints: A
case study on text categorization. Machine Learning Journal special issue on Learning in Speech
and Language Technologies. 60(1-3). pp. 169-194. Springer SBM. 2005.
30. Tadayoshi Hara, Yusuke Miyao and Jun'ichi Tsujii. Adapting a probabilistic disambiguation model
of an HPSG parser to a new domain. In Robert Dale, Kam-Fai Wong, Jian Su, Oi Yee KwongG
(Eds.), Natural Language Processing – IJCNLP 2005. LNC3651. Springer-Verlag. pp. 199-210.
2005.
31. Daisuke Okanohara and Jun'ichi Tsujii. Assigning Polarity Scores to Reviews Using Machine
Learning Techniques. In Robert Dale, Kam-Fai Wong, Jian Su and Oi Yee Kwong (Eds.), Natural
Language Processing - IJCNLP 2005. LNCS3651. Springer-Verlag. pp. 314-325. 2005.
���������� ���������
���
32. Yuka Tateisi, Akane Yakushiji, Tomoko Ohta, and Jun’ichi Tsujii. Syntax Annotation for the
GENIA corpus. In Robert Dale, Kam-Fai Wong, Jian Su and Oi Yee Kwong (Eds.), Natural
Language Processing-IJCNLP2005. LNCS3651. Springer-Verlag. pp. 222--227. 2005.
33. Jin-Dong Kim and Jun'ichi Tsujii. Word Folding: Taking the Snapshot of Words Instead of the
Whole. In Keh-Yih Su, Jun'ichi Tsujii, Jong-Hyeok Lee and Oi Yee Kwong (Eds.), Natural
Language Processing - IJCNLP 2004. LNAI 3248. Springer-Verlag. pp. 406-415. 2005.
34. Yoshimasa Tsuruoka and Jun'ichi Tsujii. Iterative CKY Parsing for Probabilistic Context-Free
Grammars. In Keh-Yih Su, Jun'ichi Tsujii, Jong-Hyeok Lee and Oi Yee Kwong (Eds.), Natural
Language Processing - IJCNLP 2004. LNAI 3248. pp. 52-60. Springer-Verlag. 2005.
35. Takashi Ninomiya, Yusuke Miyao and Jun'ichi Tsujii. A Persistent Feature-Object Database for
Intelligent Text Archive Systems. In Keh-Yih Su, Jun'ichi Tsujii, Jong-Hyeok Lee and Oi Yee
Kwong (Eds.), Natural Language Processing - IJCNLP 2004. LNAI3248. pp. 197--205.
Springer-Verlag. 2005.
36. Yoshio Nakao. A Method for Related-passage Extraction based on Thematic Hierarchy. IPSJ
Transactions on Databases. 42(SIG 10 (TOD 11)). pp. 39--53. 2001.
37. ��Mo ,¢§¨ ,©^�ª , Ñ«¬;®¯¬789:��y�ÐÑ",óN��kl,Vol.10 No.1, pp. 27 - 45, 2003/1
38. ÷� �,°^ �,�± P0,�� Mo, HTML>²³´¬D>¯~µ¶;·¸«¹²º»>�y, _`kl)M¼�½2¬D®D$ SIG12 (TOD19), pp.23-32, 2003/10
39. &¾ ¿,�� À�,Á� �H,�� Mo, �+¬;ÂÃdÄyÅgÆ+ÇÈÉÊË´”,G
NII Journal Vol.8, pp.23-34, 2004/2
40. ß= à�, �� ��, }~ ��. �æÌÅÍÎ �?�Ϭ�dÄyÅgÐP����ÑDÒÓÔϬ�. óN��kl. 11(4). pp. 3-23. 2004.
41. ��Ê, ��¦V , Õ��, "\]: ��#¡Ö��PageRank7×ØÙÈWeb}D�>ÚÛVÌÜ]ÝÞ . ªUVZ[8pE$!£E�ß�࣠(SACSIS2004), IPSJ
Symposium Series, Vol.2004, No.6, pp.103-110, 2004.5
42. Hideo Saito, Kenjiro Taura, Takashi Chikayama: Expedite: An Operating System Extension to
Support Low-Latency Communication in Non-Dedicated Clusters. ªUVZ[8pE$!£E�ß�࣠(SACSIS2004), IPSJ Symposium Series, Vol.2004, No.6, pp.443-450, 2004.5
43. ��Êá, ��¦V , Õ��: Grid �âã>äå�FæçèdéêëØE��. ªUVZ[8pE$!£E�ß�࣠(SACSIS2004), IPSJ Symposium Series, Vol.2004, No.6,
pp.225-232, 2004.5
44. ��Ê, ��¦V , Õ��, "\]: ��#¡Ö��PageRank7×ØÙÈWeb}D�>ÚÛVÌÜ]ÝÞ . In IPSJ Transactions on Advanced Computing Systems, Vol.45,
No.SIG11(ACS7), pp.465-473,2004.10
45. Hideo Saito, Kenjiro Taura, Takashi Chikayama: Expedite: An Operating System Extension to
Support Low-Latency Communication in Non-Dedicated Clusters. In IPSJ Transactions on
Advanced Computing Systems, Vol.45, No.SIG11(ACS7), pp.229-237,2004.10
46. Minoru Yoshida, Kentaro Torisawa, and Jun'ichi Tsujii.G Integrating Tables on the World Wide Web,
<JKL)M¼�½(Transactions of the Japanese Society for Artificial Intelligence). 19(6). pp.
548-560, 2004/10
47. ì�í´, ��¦V , Õ��2îïðäåZ[déêëØóñVÌïð.K˯. ªUV
Z[8pE$!£E�ß�࣠(SACSIS2005),IPSJ Symposium Series, Vol.2005, No.5,
pp.311-319,2005.5
48. ì�í´, ��¦V , Õ��2îïðäåZ[déêëØóñVÌïð.K˯. In IPSJ
Transactions on Advanced Computing Systems¦vol.46, No.SIG12(ACS11)¦ pp.236-244¦2005
49. ��Mo2 "Kiwi:ò��yz./E$!£" óô�õ_`kluv 6n, pp.116-123, 2005/10
�
���������� ����������� ��
���
�*������Y��4���kK����F4���_�
�D���$H��� � Y�F� �¡ �H��� ¢£ �_�
1. Naoki Yoshinaga, Yusuke Miyao, Kentaro Torisawa and Jun'ichi Tsujii. Resource sharing among
HPSG and LTAG communities by a method of grammar conversion from FB-LTAG to HPSG. In
the Proceedings of ACL/EACL Workshop on Sharing Tools and Resources for Research and
Education. pp. 39--46. Morgan Kaufman Publishers. 2001.
2. Yoshio Nakao. How small a distinction among summaries can an IR-based evaluation method
identify? In the Proceedings of NAACL 2001 Workshop on Automatic Summarization
(WAS2001). pp. 69--78. 2001.
3. Kenji Nishida, Kentaro Torisawa and Jun'ichi Tsujii. Compiling an HPSG-based grammar into
more than one CFG. In the Proceedings of PACLING 2001. pp. 199--206. 2001.
4. Naoki Yoshinaga, Yusuke Miyao, Kentaro Torisawa and Jun'ichi Tsujii. Efficient LTAG parsing
using HPSG parsers. In the Proceedings of Pacific Association for Computational Linguistics
(PACLING 2001). pp. 342--351. 2001.
5. Hideki Mima, Sophia Ananiadou and Goran Nenadic. Improving Knowledge Acquisition Through
Automatic Term Recognition. In the Proceedings of Panhellenic Conference on Human Computer
Interaction (PC-HCI 2001). pp. 177-182. 2001.
6. Tomoko Ohta, Yuka Tateisi, Jin-Dong Kim, Hideki Mima and Jun'ichi Tsujii. Ontology Based
Corpus Annotation and Tools. In the Proceedings of the 12th Genome Informatics 2001. pp.
469--470. 2001.
7. Yoshio Nakao. How small a distinction among summaries can the evaluation method identify?
In the Proceedings of the NTCIR-2 workshop. pp. 341--348. 2001.
8. Minoru Yoshida, Kentaro Torisawa and Jun'ichi Tsujii. Extracting ontologies from World Wide
Web via HTML tables. In the Proceedings of the Pacific Association for Computational
Linguistics (PACLING 2001). pp. 332-341. 2001.
9. Jun'ichi Kazama, Yusuke Miyao and Jun'ichi Tsujii. A Maximum Entropy Tagger with
Unsupervised Hidden Markov Models. In the Proceedings of the Sixth Natural Language
Processing Pacific Rim Symposium. pp. 333--340. 2001.
10. Minoru Yoshida, Kentaro Torisawa and Jun'ichi Tsujii. A method to integrate tables of the World
Wide Web. In the Proceedings of the first International Workshop on Web Document Analysis
(WDA 2001). pp. 31-34. 2001.
11. Jin-Dong Kim, Tomoko Ohta, Yuka Tateisi, Hideki Mima and Jun'ichi Tsujii. XML-Based
Linguistic Annotation of Corpus. In the Proceedings of the first NLP and XML Workshop held at
NLPRS 2001. pp. 47--53. 2001.
12. Naoki Yoshinaga and Yusuke Miyao. Grammar conversion from LTAG to HPSG. In the
Proceedings of the sixth ESSLLI Student Session. pp. 309--324. 2001.
13. Tomoko Ohta, Yuka Tateisi and Jun'ichi Tsujii. Tools for Ontology-based Corpus Annotation. In
the Proceedings of the sixth Pacific Symposium on Biocomputing (PSB 2001). pp. 112. 2001.
14. Akane Yakushiji, Yuka Tateisi, Yusuke Miyao and Jun'ichi Tsujii. Event extraction from
biomedical papers using a full parser. In the Proceedings of the sixth Pacific Symposium on
Biocomputing (PSB 2001). pp. 408-419. 2001.
15. Takaki Makino, Kazuyuki Aihara and Jun'ichi Tsujii. Towards Sentence Understanding: Phase
Arbitration in Temporal-Coding Memory Mechanism. In the The Second Workshop on Natural
Language Processing and Neural Networks (NLPNN 2001). pp. 46--52. 2001.
16. Yusuke Miyao and Jun'ichi Tsujii. Maximum Entropy Estimation for Feature Forests. In the
Proceedings of Human Language Technology Conference (HLT 2002). 2002.
17. Hideki Mima, Sohia Ananiadou, Goran Nenadic and Jun'ichi Tsujii. XML Tag Information
Management System: A Workbench for Ontology-based Knowledge Acquisition and Integration.
In the Proceedings of Human Language Technology Conference (HLT 2002). 2002.
���������� ���������
���
18. Hideki Mima, Sophia Ananiadou, Goran Nenadic and Jun'ichi Tsujii. TIMS - A Workbench for
Ontology-based Knowledge Acquisition and Integration. In the Proceedings of Natural Language
Processing in Biomedical Applications (NLPBA 2002). 2002.
19. Takashi Ninomiya, Yusuke Miyao and Jun'ichi Tsujii. Lenient Default Unification for Robust
Processing within Unification Based Grammar Formalisms. In the Proceedings of the 19th
International Conference on Computational Linguistics (COLING 2002). pp. 744--750. 2002.
20. Takashi Ninomiya, Takaki Makino and Jun'ichi Tsujii. An Indexing Scheme for Typed Feature
Structures. In the Proceedings of the 19th International Conference on Computational Linguistics
(COLING 2002). pp. 1248-1252. 2002.
21. Junko Araki, Takashi Ninomiya, Takaki Makino and Jun'ichi Tsujii. Action Vectors for
Interpreting Route Descriptions. In the Proceedings of the AAAI-02 Workshop on Spatial and
Temporal Reasoning. 2002.
22. Minoru Yoshida. Extracting Attributes and Their Values from Web Pages. In the Proceedings of
the ACL 2002 Student Research Workshop. pp. 72--77. 2002.
23. Tomoko Ohta, Yuka Tateisi, Hideki Mima and Jun'ichi Tsujii. GENIA Corpus: an Annotated
Research Abstract Corpus in Molecular Biology Domain. In the Proceedings of the Human
Language Technology Conference (HLT 2002). 2002.
24. Jun'ichi Kazama, Takaki Makino, Yoshihiro Ohta and Jun'ichi Tsujii. Tuning Support Vector
Machines for Biomedical Named Entity Recognition. In the Proceedings of the Natural Language
Processing in the Biomedical Domain (ACL 2002). pp. 1-8. 2002.
25. Jin-Dong Kim and Jun'ichi Tsujii. Corpus-Based Approach to Biological Entity Recognition. In
the Proceedings of the Second Meeting of the Special Interest Group on Text Data Mining of ISMB
2002. 2002.
26. Tadayoshi Hara, Yusuke Miyao and Jun'ichi Tsujii. Clustering for obtaining syntactic classes of
words from automatically extracted LTAG grammars. In the Proceedings of the sixth International
Workshop on Tree Adjoining Grammars and Related Frameworks (TAG+6). pp. 227-233. 2002.
27. Junko Araki, Takashi Ninomiya, Takaki Makino and Jun'ichi Tsujii. Two Perspective systems
Using a Route as a Reference Object. In the Proceedings of the sixth World Multiconference on
Systemics, Cybernetics and Informatics (SCI 2002). 2002.
28. Naoki Yoshinaga, Yusuke Miyao and Jun'ichi Tsujii. A formal proof of strong equivalence for a
grammar conversion from LTAG to HPSG-style. In the Proceedings of the sixth International
Workshop on Tree Adjoining Grammars and Related Frameworks (TAG+6). pp. 187--192.
2002.
29. M. Yoshida, Extracting Attributes and Their Values from Web PagesIn Proceedings of the ACL-02
Student Research Workshop, pages 72-77, 2002/7
30. Hiroshi Nakagawa and Tataunori Mori, A Simple but Powerful automatic Term Extraction Method,
Computerm2: 2nd International Workshop on Computational Terminology, COLING-2002
WORKSHOP, pp. 29-35,Taipei, 2002/8
31. Takashi Masuyama and Hiroshi Nakagawa,Applying Cascaded Feature Selection to SVM Text
Categorization", 3rd International Workshop on Natural Language and Information Systems, pp.
241-245, Aix-en-Provence, France, 2002/9
32. Yoshiaki Kawasaki, Jun'ichi Kazama and Jun'ichi Tsujii. Extracting Biomedical Ontology from
Textbooks and Article Abstracts. In the Proceedings of the SIGIR'03 Workshop on Text Analysis
and Search for Bioinformatics. pp. 44-50. 2003.
33. Akane Yakushiji, Yuka Tateisi, Yusuke Miyao, Naoki Yoshinaga and Jun'ichi Tsujii. A Debug
Tool for Practical Grammar Development. In the Companion Volume to the Proceedings of 41st
Annual Meeting of the Association for Computational Linguistics. pp. 173--176. 2003.
34. Jun-ichi Tsujii. New Perspectives of Linguistic Study. In the ISCA and IEEE workshop on
Spontaneous Speech Processing and Recognition. pp. 151-157. 2003.
���������� ����������� ��
���
35. Katsuya Masuda, Takashi Ninomiya, Yusuke Miyao, Tomoko Ohta and Jun'ichi Tsujii. A Robust
Retrieval Engine for Proximal and Structural Search. In the Proceedings of HLT-NAACL 2003
Short papers. pp. 58--60. 2003.
36. Yusuke Miyao, Takashi Ninomiya and Jun'ichi Tsujii. Lexicalized Grammar Acquisition. In the
Proceedings of the 10th Conference of the European Chapter of the Association for Computational
Linguistics (EACL) companion volume. pp. 127--130. 2003.
37. Jun'ichi Kazama and Jun'ichi Tsujii. Evaluation and Extension of Maximum Entropy Models with
Inequality Constraints. In the Proceedings of the 2003 Conference on Empirical Methods in
Natural Language Processing (EMNLP 2003). pp. 137-144. 2003.
38. Yoshimasa Tsuruoka and Jun'ichi Tsujii. Probabilistic Term Variant Generator for Biomedical
Terms. In the Proceedings of the 26th Annual International ACM SIGIR Conference. pp.
167--173. 2003.
39. Naoki Yoshinaga, Kentaro Torisawa and Jun'ichi Tsujii. Comparison between CFG filtering
techniques for LTAG and HPSG. In the Proceedings of the 41st ACL companion volume. pp.
185--188. 2003.
40. Jin-Dong Kim, Hae-Chang Rim and Jun'ichi Tsujii. Self-Organizing Markov Models and Their
Application to Part-of-Speech Tagging. In the Proceedings of the 41st Annual Meeting of the
Association for Computational Linguistics. pp. 296-302. 2003.
41. Katsuya Masuda. A Ranking Model of Proximal and Structural Text Retrieval Based on Region
Algebra. In the Proceedings of the ACL 2003 Student Research Workshop. pp. 50--57. 2003.
42. Yoshimasa Tsuruoka and Jun'ichi Tsujii. Boosting Precision and Recall of Dictionary-Based
Protein Name Recognition. In the Proceedings of the ACL-03 Workshop on Natural Language
Processing in Biomedicine. pp. 41-48. 2003.
43. Zhonghua Yu, Yoshimasa Tsuruoka and Jun'ichi Tsujii. Automatic Resolution of Ambiguous
Abbreviations in Biomedical Texts using Support Vector Machines and One Sense Per Discourse
Hypothesis. In the Proceedings of the SIGIR'03 Workshop on Text Analysis and Search for
Bioinformatics. pp. 57-62. 2003.
44. Yusuke Miyao and Jun'ichi Tsujii. A model of syntactic disambiguation based on lexicalized
grammars. In the Proceedings of the Seventh Conference on Natural Language Learning (CoNLL)
at HLT-NAACL 2003. pp. 1--8. 2003.
45. Takuya Matsuzaki, Yusuke Miyao and Jun'ichi Tsujii. An Efficient Clustering Algorithm for
Class-based Language Models. In the Proceedings of the Seventh Conference on Natural
Language Learning (CoNLL) at HLT-NAACL 2003. pp. 119--126. 2003.
46. Yoshimasa Tsuruoka and Jun'ichi Tsujii. Training a Naive Bayes Classifier via the EM Algorithm
with a Class Distribution Constraint. In the Proceedings of the Seventh Conference on Natural
Language Learning (CoNLL) at HLT-NAACL 2003. pp. 127--134. 2003.
47. Yusuke Miyao, Takashi Ninomiya and Jun'ichi Tsujii. Probabilistic modeling of argument
structures including non-local dependencies. In the the Proceedings of the Conference on Recent
Advances in Natural Language Processing (RANLP) 2003. pp. 285--291. 2003.
48. Takashi Masuyama and Hiroshi Nakagawa, Cascaded Feature Selection in SVM Text Categorization,
Lecture Note in Computer Science: the proceedings of 4th International Coference on Intelligent
Text Processing and Computational Linguistics CICLing-2003, pp.588-591, Mexico City, 2003/2
49. Kumiko Tanaka-Ishii, Michiko Abe, and Hiroshi Nakagawa., Categorization of movies using
comments, Proceedings of PACLING'03 (Pacific Association for Computational LINGuistics),
pp.221-229, Halifax, Nova Scotia, Canada, 2003/8 (Best Paper Award of PACLING'03)
50. Hiroko Nakanishi, Yusuke Miyao and Jun'ichi Tsujii. Using Inverse Lexical Rules to Acquire a
Wide-coverage Lexicalized Grammar. In the IJCNLP 2004 Workshop on Beyond Shallow
Analyses. 2004.
���������� ���������
���
51. Yuka Tateisi, Ohta, Tomoko and Tsujii, Jun-ichi. Annotation of Predicate-argument Structure of
Molecular Biology Text. In the IJCNLP-04 workshop on Beyond Shallow Analyses. 2004.
52. Jun-ichi Tsujii. Thesaurus or logical ontology, which do we need for mining text? (keynote
speech). In the Proc. of Language resources and evaluation conference (LREC 2004). Vol.III. pp.
pp IX-XVI. 2004.
53. Yuka Tateisi and Jun'ichi Tsujii. Part-of-Speech Annotation of Biology Research Abstracts. In
the Proceedings of 4th International Conference on Language Resource and Evaluation (LREC2004).
IV. pp. 1267-1270. 2004.
54. Yusuke Miyao and Jun'ichi Tsujii. Deep Linguistic Analysis for the Accurate Identification of
Predicate-Argument Relations. In the Proceedings of COLING 2004. pp. 1392-1397. 2004.
55. Yoshimasa Tsuruoka, Yusuke Miyao and Jun'ichi Tsujii. Towards efficient probabilistic HPSG
parsing: integrating semantic and syntactic preference to guide the parsing. In the Proceedings of
IJCNLP-04 Workshop: Beyond shallow analyses - Formalisms and statistical modeling for deep
analyses. 2004.
56. Naoki Yoshinaga and Jun'ichi Tsujii. Generalizing Subcategorization Frames Acquired from
Corpora Using Lexicalized Grammars. In the Proceedings of TAG+7. pp. 104--110. 2004.
57. Kenta Oouchida, Naoki Yoshinaga and Jun'ichi Tsujii. Context-free Approximation of LTAG
towards CFG Filtering. In the Proceedings of TAG+7. pp. 171--177. 2004.
58. Jin-Dong Kim and Jun'ichi Tsujii. Word Folding: Taking the Snapshot of Words Instead of the
Whole. In the Proceedings of the First International Joint Conference on Natural Language
Processing. pp. 61--68. 2004.
59. Jin-Dong Kim, Tomoko Ohta, Yoshimasa Tsuruoka, Yuka Tateisi and Nigel Collier. Introduction
to the Bio-Entity Recognition Task at JNLPBA. In the Proceedings of the International Workshop
on Natural Language Processing in Biomedicine and its Applications (JNLPBA-04). pp. 70--75.
2004.
60. Hiroko Nakanishi, Yusuke Miyao and Jun'ichi Tsujii. An Empirical Investigation of the Effect of
Lexical Rules on Parsing with a Treebank Grammar. In the Proceedings of the third TLT2004.
pp. 103--114. 2004.
61. Naoki Yoshinaga. Improving the Accuracy of Subcategorizations Acquired from Corpora. In the
The 42nd ACL Student Session. pp. 43--48. 2004.
62. Akane Yakushiji, Yuka Tateisi, Yusuke Miyao and Jun'ichi Tsujii. Finding Anchor Verbs for
Biomedical IE Using Predicate-Argument Structures. In the the Companion Volume to the
Proceedings of 42st Annual Meeting of the Association for Computational Linguistics. pp.
157--160. 2004.
63. Minoru Yoshida and Hiroshi Nakagawa. Specification Retrieval -- How to Find Attribute-Value
Information on the Web?, Proceedings of IJCNLP (International Joint Conference of Natural
Language Processing) 2004, pp.520-527, Hainan-Island, China, March 2004/3
64. Koichi Yamada, Hisashi Komine, Hiroshi Kinukawa and Hiroshi Nakagawa, Abstract of Abstract: A
New Summarizing Method based on Document Frequency and Clause Length, SCI2004(The 8th
World Multi-Conference on Systemics, Cybernetics and Informatics), Volume XIV, pp.56-61,
Orland,U.S.A., 2004/7
65. Takeshi Masuyama, Satoshi Sekine and Hiroshi Nakagawa, Automatic Construction of Japanese
KATAKANA Variant List from Large Corpus, COLING2004(Proceedings of the 20th International
Conference on Computational Linguistics), pp. 1214-1219, Geneva, Switzerland, 2004/8
66. Hiroko Nakanishi, Yusuke Miyao and Jun'ichi Tsujii. Probabilistic models for disambiguation of
an HPSG-based chart generator. In the Proc. of IWPT 2005. 2005.
67. Takashi Ninomiya, Yoshimasa Tsuruoka, Yusuke Miyao and Jun'ichi Tsujii. Efficacy of Beam
Thresholding, Unification Filtering and Hybrid Parsing in Probabilistic HPSG Parsing. In the Proc.
of IWPT 2005. 2005.
���������� ����������� ��
���
68. Yusuke Miyao and Jun'ichi Tsujii. Probabilistic disambiguation models for wide-coverage HPSG
parsing. In the Proceedings of ACL 2005. pp. 83-90. 2005.
69. Yoshimasa Tsuruoka and Jun'ichi Tsujii. Bidirectional Inference with the Easiest-First Strategy for
Tagging Sequence Data. In the Proceedings of HLT/EMNLP 2005. pp. 467-474. 2005.
70. Yoshimasa Tsuruoka, Yuka Tateishi, Jin-Dong Kim, Tomoko Ohta, John McNaught, Sophia
Ananiadou and Jun'ichi Tsujii. Developing a Robust Part-of-Speech Tagger for Biomedical Text.
Advances in Informatics - 10th Panhellenic Conference on Informatics, LNCS 3746, pp. 382-392.
2005.
71. Takuya Matsuzaki, Yusuke Miyao and Jun'ichi Tsujii. Probabilistic CFG with Latent Annotations.
In the Proceedings of the 43rd Annual Meeting of the Association for Computational Linguistics.
pp. 75-82. 2005.
72. Yoshimasa Tsuruoka and Jun'ichi Tsujii. Chunk Parsing Revisited. In the Proceedings of the 9th
International Workshop on Parsing Technologies (IWPT 2005). pp. 133-140. 2005.
73. Yoshimasa Tsuruoka, Sophia Ananiadou and Jun'ichi Tsujii. A Machine Learning Approach to
Acronym Generation. In the Proceedings of the ACL-ISMB Workshop on Linking Biological
Literature, Ontologies and Databases: Mining Biological Semantics. pp. 25-31. 2005.
74. Akane Yakushiji, Yusuke Miyao, Yuka Tateisi and Jun'ichi Tsujii. Biomedical Information
Extraction with Predicate-Argument Structure Patterns. In the the Proceedings of the First
International Symposium on Semantic Mining in Biomedicine. pp. 60--69. 2005.
75. Kumiko Tanaka-Ishii and Hiroshi Nakagawa, A Multilingual Usage Consultation Tool based on
Internet Searching ---More than search engine, Less than QA, The 14th International World Wide
Web Conference (WWW2005), Chiba, Japan, pp.363-371. 2005/5 (Best Presentation Awards of
WWW2005)
76. Ayako Hoshino and Hiroshi Nakagawa, A real-time multiple-choice question generation for
language testing -- a preliminary study--, 43rd ACL2005 Second Workshop on Building Educational
Applications Using Natural Language Processing, pp.17-20. Ann Arbor, 2005/7
77. Takeshi Masuyama and Hiroshi Nakagawa, Web-based Acquisition of Japanese Katakana Variants,
SIGIR2005 (The 28th Annual International ACM SIGIR Conference) . pp.338-344, Salvador, Brazil,
2005/8
78. Minoru Yoshida, Hiroshi Nakagawa, "Automatic Term Extraction based on Perplexity of Compound
Words", IJCNLP'05 (The 2'nd International Joint Conference of Natural Language Processing,
pp.269--279) Juje, Korea, 2005/10
79. âî D��, }~ ��, ¤¥ ¦§�. ö÷�èr>Iøùúû>ÐÑE$!£. ��
kl)MH 7I�VPMü²¼�]. pp. 38--41. 2001.
80. áâ ã´, �� ��, ¤¥ ¦§�, }~ ��. LTAG� 7×Ø HPSGÞDýdyþg¯�ýþ. ��kl)MH 7I�VPMü²¼�]. pp. 277--280. 2001.
81. µE �, ß= à�, �� ��, }~ ��. òúÌ_`�dyþgÍÎ �?�Ϭ�7×Ø�)/<�)¼�>��y�µ¶�. ��kl)MH 8 I�VPMü²¼�]. pp.
21--24. 2002.
82. ÷� øù, Î� Ï, }~��. GCL �þX���7×د~Y!ú$�>�#./.
��kl)MH 8I�VPMü²¼�]. pp. 459--462. 2002.
83. ñò àÝ, Î� Ï, �î �´, }~��. �>Ë�����ëØ���º>���.
H 16I<JKL)MNOPM. 2002.
84. &¾¿,Á��H,��Mo,��>�+¬;�>Ã�dÄyÅgôõÉÊË´, _`kl
)M NLuvM NL-149, pp.73-80, 2002/5
85. ���Ý,��S±Ý,��Mo,?Ö��dyþg��>S , _`kl)M NL uvM
NL-150, pp.105-110, 2002/7
���������� ���������
���
86. °^ �,÷� �,��Mo, HTML >²³´¬D>��;·¸«¹²º»>�y, _`
kl)MóN��kluvM NL-151, pp.35-42, 2002/9
87. �^�<,��S±Ý,��Mo, web./789:ò��5V KWIC, _`kl)M NLuv
M NL-152, pp.115-122, 2002/11
88. °^ �,÷� �,��Mo,"·¸«¹²ºdx�ÅgHTML>²³´¬D>¯~µ¶;��", _`)E�ß�à£,pp.5-8,_`kl)M_`)8!uvM"#, 2003/1
89. �^�<, ��S±Ý, ��Mo, ./����789:ò��yz�$vD�:Kiwi",��
kl)MH 9IPMü²%n A1-5, 2003.4,¤�³ü²&¥
90. á� ��, ÇÈ ÉÊ, ����, }~��. WÛV��Ϭ�7'ÔØÞÖD>WÛS(d)/ëØ� ;*>�y. ��kl)MH 9 I�VPM¼�]. pp. 250--253.
2003.
91. +� È�, �� ��, }~��. ¯�ýþ7'ÔØ�,qVÌ-./>¯�VZ[. �
�kl)MH 9I�VPM¼�]. pp. 549--552. 2003.
92. P�� �§, áâ ã´, Î�G Ï, �� ��, }~ ��. �0Y� 7'ÔØ�0wx>¯~V\1789:ó5S . ��kl)MH 9 I�VPM¼�]. pp. 553--556.
2003.
93. åæç eèé, »¼ ½¾, �� ��, áâ ã´, }~ ��. �yVÌ� d2üëØg3>¬4�ÒvD�. _`kl)Muv`5 NL-155. pp. 19--24. 2003.
94. �� ��, Î� Ï, �� ��, }~ ��. _`./>g3>�Ñ�#6�7��#$¡�Ò. ��kl)MH 10I�VPMü²¼�]. pp. 21-24. 2004.
95. �� u8, ÇÈ ÉÊ, }~ ��. Y9#$¡�ÒdyþgWeb�+èr>Ç:;<>=t. ��kl)MH 10I�VPMü²¼�]. pp. 333-336. 2004.
96. P�� �§, áâ ã´, }~ ��. CFG�"�¡�Òdyþg"ÚÛÌ LTAGÞDý>¯°. ��kl)MH 10I�VPMü²¼�]. pp. 361--364. 2004.
97. ÷� øù, Î� Ï, �� ��, }~ ��. ¶·�dyþg¯~Y!ú$�./>>¦ã$CDæ�ÌϬ�. ��kl)MH 10 I�VPMü²¼�]. pp. 377--380.
2004.
98. } ¡, á�|, ��Mo, �0?=>¯~789:��y�ÐÑ_`kl)M NL uvM
159, pp.155-162, 2004/1
99. @îA�, ¢§¨, òBCÝ, ��Mo,DEFÔ> 2�Ò�;G,'DdÄyÅg3�²«ÐÑ": _`kl)M NLuvM 159, pp.169-176, 2004/1
100. á�|, ��Mo, HHHèr>r�Ir�/_`}D�>./, ��kl)MH 10I�VPMü²¼�], pp. 620—623, 2004/3
101. } ¡, á�|, ��Mo, �ZV'×JÒ�VK�dyþg��y�ÐÑ��kl)MH 10I�VPMü²¼�], pp. 33—36, 2004/3
102. @îA�, ¢§¨, òBCÝ, ��Mo,"DEFÔ>L�Ò�;G,'DdÄyÅg3�²«ÐÑ", ��kl)M H 10I�VPM, A1-1, 2004/3
103. } ¡, á�|, ��Mo, �ZV'×JÒ�VK�dyþg��y�ÐÑ", ��kl)
M H 10I�VPM, A1-6, 2004/3
104. &¾¿, ��À�, Á��H, ��Mo, ÙÈÂÐÑ7ÈÊ7'ÔØMNOë��@Ë´>.�, ��kl)M H 10I�VPM, C2-5, 2004/3
105. �^�<, ��S±Ý, ��Mo, ò��yz�$vD�:Kiwi>�PV-., ��kl)M
H 10I�VPM, D5-1, 2004/3
106. á�|, ��Mo, HHHèr>r�Ir�/_`}D�>./", ��kl)M H 10I�
VPM, D5-3, 2004/3
107. ¢�£,} ¡,á�|,��Mo, ��y�ÐÑË >!$�?Q#EF�RS�_`kl)Muv`5 ^óN��kljNo.161, 2004/5
���������� ����������� ��
���
108. ��¦V : Grid Explorer: A Tool for Discovering, selecting, and using Distributed Resources
Efficiently. 2004�äå/ST/T4kl7'ëØ^U¢jV DIWD#EF�X(SWoPP U
¢ 2004), _`kl)Muv`5 2004-HPC-99, Vol.2004, No.81, pp.235-240, 2004.8
109. Y�*Î, ��¦V , Qh�*: ¯Zd5V7[\ëØÖ�qD�Þ�E�ÒE$!£.
2004�äå/ST/T4kl7'ëØ̂ U¢jV DIWD#EF�X(SWoPP U¢ 2004), _
`kl)Muv`5 2004-HPC-99, Vol.2004, No.81, pp.241-246, 2004.8
110. ]o�^�, ��¦V , Õ��: ¼l´>_`aL��ø>äåY7'ÔØ ClauseGb>
Úc7dþe. 2004�äå/ST/T4kl7'ëØ^U¢jV DIWD#EF�X(SWoPP
U¢ 2004), fÝ_`�)Mhiuv`5, Vol.104, No.240, pp.25-30, 2004.8
111. ghi�, ��¦V , Õ��: �{RSýþ789:jó5äåY>ÚÛV��. 2004
�äå/ST/T4kl7'ëØ^U¢jV DIWD#EF�X(SWoPP U¢ 2004), _`
kl)M¼�½: X�Ò@�Ò, Vol.46, No.SIG 1(PRO 24),2004.8
112. )�³#, ��¦V , Õ��: Phoenix X�Ò@�ÒϬ�y>]k�. 2004 �äå/
ST/T4kl7'ëØ^U¢jV DIWD#EF�X(SWoPP U¢ 2004), _`kl)M
uv`5 2004-HPC-99, Vol.2004, No.81, pp.223-228, 2004.8
113. ì�í´, ��¦V , Õ��: Phoenix X�Ò@�ÒϬ�7'ÔØïð.K�æ¡. 2004�äå/ST/T4kl7'ëØ̂ U¢jV DIWD#EF�X(SWoPP U¢ 2004),
_`kl)Muv`5 2004-HPC-99, Vol.2004, No.81, pp.229-234. 2004.8
114. P�ÜÝ, »¼½¾, YU�, åæçeèé, }~��. <Ia)Sî>ÒÓl?DÞ$2GENIA?DÞ$>{Z;6�. ��kl)MH 11I�VPMü²¼�]. pp. 506--509
2005.
115. åæçeèé, ����, »¼½¾, }~��. m�w¯~ÞD�dyþg�)I<�)Sî_`ÐÑ. ��kl)MH 11I�VPMü²¼�]. pp. 93--96. 2005.
116. -�à�, }~��. DEFÔ'D;�þ�n'DdyþgopÌq^�./. ��kl
)MH 11I�VPMü²¼�]. pp. 568--571. 2005.
117. Èî�P0, }~��. -.�7rëØÎs�t>ó5Óu. ��kl)MH 11 I�VPMü²¼�]. pp. 664-667. 2005.
118. ¢�£,á�|, ��Mo, "vX�ÞDXQúE!"789:ÙÈ�ÐÑ >uv", ��k
l)M H 11I�VPM, B2-8, 2005/3
119. ÷�w�, ��Mo, "Web¬DdÄyÅgxxyz²{>ó5st", ��kl)M H
11I�VPM, B2-9, 2005/3
120. �^k| ,á�| , ��Mo , "�Dx�?DÞ$èr>!ú$� �B�ÒvD�2PortableKiwi", ��kl)M H 11I�VPM, C1-8, 2005/3
121. ]o�^�,��¦V ,Õ��2¼l´>_`aL��ø7'ÔØ�>RS'D789:ÚÛVÌ�}/�~.2005 �äå/ST/T4kl7'ëØ^�#jV DIWD#EF�X(SWoPP �# 2005),_`kl)Muv`5 2005-HPC-103,pp.91-96,2005.8
122. )�³#,��¦V ,Õ��2R·Ö�qD�Þ�E�ÒE$!£y>��d��Åg¯Zxl. 2005�äå/ST/T4kl7'ëØ̂ �#jV DIWD#EF�X(SWoPP �# 2005),
_`kl)Muv`5 2005-HPC-103,pp.181-185,2005.8
123. ì�í´,��¦V ,Õ��2vVæ����â7'ÔØóñVÌïð.K˯. 2005�
äå/ST/T4kl7'ëØ^�#jV DIWD#EF�X(SWoPP �# 2005),_`kl
)Muv`5 2005-OS-100,pp. 93—100,2005.8
124. a�bc,��¦V 2"þî���d�d�à$�� . 2005�äå/ST/T4kl7'ëØ^�#jV DIWD#EF�X (SWoPP �# 2005),_`kl)Muv`5
2005-HPC-103,pp. 121—126,2005.8
125. �~�ù,a�bc,��¦V ,Õ��2"þ�DX�yÛ>�;ã"�Ì��#¡Ö��GC. 2005�äå/ST/T4kl7'ëØ̂ �#jV DIWD#EF�X(SWoPP �# 2005),
_`kl)MuvM�� PRO-2005-2,2005.8 (�y��>Nã.�]ÌÅ)
���������� ���������
���
126. "\�,��¦V ,Õ��2 �ÒQDEF�déêëØST]XÆæ��#�. 2005�äå/ST/T4kl7'ëØ^�#jV DIWD#EF�X(SWoPP �# 2005),_`kl)
MuvM�� PRO-2005-2,2005.8 (�y��>Nã.�]ÌÅ)
127. ��¦V 2Grid/P2P X�Ò@�ÒϬ�;'®hi. PPL Summer SchoolPPL Summer
School 2005,�¨P),2005.9 (��'�>Nã.�]ÌÅ)
�
����¤f�<��� � Y�F� ¥ �H��� D¦ �_�
1. Tomoko Ohta, Yuka Tateisi, Jin-Dong Kim, Sang-Zoo Lee and Jun'ichi Tsujii. GENIA corpus: A
Semantically Annotated Corpus in Molecular Biology Domain. In the Proceedings of the ninth
International Conference on Intelligent Systems for Molecular Biology (ISMB 2001) poster session.
pp. 68. 2001.
2. Tomoko Ohta, Yuka Tateisi, Jin-Dong Kim and Jun'ichi Tsujii. The GENIA Corpus: an Annotated
Corpus in Molecular Biology Domain. In the Proceedings of the 10th International Conference on
Intelligent Systems for Molecular Biology (ISMB 2002) poster session. 2002.
3. Kumiko Tanaka-Ishii, Masato Yamamoto, and Hiroshi Nakagawa, Kiwi: A Multilingual Usage
Consultation Tool based on Internet Searching, Proceedings of the Interactive
Posters/Demonstrations, ACL-03, pp.105-108, Sapporo, 2003/7
4. Hong-Woo Chun, Tomoko Ohta, Jin-Dong Kim and Jun'ichi Tsujii. Building Patterns for
Biomedical Event Extraction. In the The 15th International conference on Genome Informatics
(GIW)-Poster and Software Demonstrations. 15(2). pp. 163. Universal Academy Press, INC.
2004.
5. Miura Kenji, Yoshimasa Tsuruoka and Jun'ichi Tsujii. Automatic acquisition of concept relations
from web documents with sense clustering. In the IJCNLP 2004 Interactive Poster/Demo.
pp.37-40. 2004.
6. Hidetaka Masuda, Shuuich Tsukamoto and Hiroshi Nakagawa, Recognition of HTML Table
Structure, Proceedings of IJCNLP (International Joint Conference of Natural Language Processing)
2004, pp.183-188, Hainan-Island, China, March 2004/3
7. Hiroshi Nakagawa, Hiroyuki Kojima, and ra Maeda, Cinese Term Extraction from Web Pages Based
on Compound word Productivity, 42nd Annual Meeting of the Association for Computational
Linguistics (ACL2004), Third SIGHAN Workshop on Chinese Language Processing, pp.79-85,
Barcelona, Spain, 2004/7
8. Minoru Yoshida and Hiroshi Nakagawa, Reformatting Web Documents via Header Trees, 43rd
ACL2005 Poster/Demo Session, pp.121-124. Ann Arbor, 2005/7
9. Ayako Hoshino and Hiroshi Nakagawa, WebExperimenter for Multiple Choice Question
Generation", HLT-EMNLP-05(Human Language Technology Conference Conference on Empirical
Methods in Natural Language Processing, pp.18-19), Demo session, Vancouver, B.C., Canada
2005/10
10. Okanohara Daisuke. Partially Decodable Compression with Static PPM. In the Data
Compression Conference 2005 poster session. 2005.
11. ì�í´, ��¦V , Õ��: ST�â7'ÔØîïðäåZ[déêëØ��æ¡. ªUVZ[8pE$!£E�ß�࣠(SACSIS2004), May 2004.
12. ]o�^�, ��¦V , Õ��2v#$�âã>äå SAT-solver >�. ªUVZ[8pE$!£E�ß�࣠(SACSIS2005),May 2005.
13. )�³#, ��¦V , Õ��25V$ÞB�Òv¡DdyþgR·Ö�qD�Þ�E�ÒE$!£y>]X�. ªUVZ[8pE$!£E�ß�࣠(SACSIS2005),May 2005.
14. �~�ù, a�bc, ��¦V , Õ��2"þ�DX�yÛ>�;ã"�Ì��#¡Ö�� GC. ªUVZ[8pE$!£E�ß�࣠(SACSIS2005),May 2005.
15. "\�, ��¦V , Õ��2 �ÒQDEF�déêëØST]XÆæ��#�. ªUV
Z[8pE$!£E�ß�࣠(SACSIS2005),May 2005.
���������� ����������� ��
���
16. a�bc, ��¦V 2"þî���d�d�à$�� . ªUVZ[8pE$!£E�ß�࣠(SACSIS2005),May 2005.
�*)�§¨�©Y�F� � ¦ �H��� � �Hª�«��¦��_�
�������-��
�
�*¬�®¯��
�D��®�1. ��kl)MH 9MPM �³ü²&� 2003.4�
�^�<� ��S±Ý� ��Mo� "./����789:ò��yz�$vD�:Kiwi"�
ü²%n A1-5
2. Best Paper Award of PACLING'03� Aug. 2003�
Kumiko Tanaka-Ishii� Michiko Abe� Hiroshi Nakagawa. "Categorization of movies using
comments"� PACLING'03 (Pacific Association for Computational LINGuistics)� Halifax� Nova
Scotia� Canada� Aug. 2003
3. ��Ê� ��¦V � Õ��� "\]H2��#¡Ö�� PageRank7×ØÙÈWeb
}D�>ÚÛVÌÜ]ÝÞG SACSIS 2004 )<�³¼�&G 2004� 5�
4. P� Adrian&: O Royal Academy, 2004
���$DP);��P)�}~uv��>^<Ia)>g3>!ú$� �B�Òhij>G3uv7rÅe�L�������G
G
5. ì�í´���¦V �Õ��2îïðäåZ[déêëØóñVÌïð.K˯G SACSIS
2005G �³L�uv&G 2005� 5�
6. Best Presentation Awards of WWW2005�
Kumiko Tanaka-Ishii (=presentor)� Hiroshi Nakagawa (= co-author)� "A Multilingual Usage
Consultation Tool based on Internet Searching ? More than search engine� Less than QA ?"� Chiba�
Japan� May 2005
7. Eclipse Innovation Grant(EIG)2QO���
��P)¤}~uv�¥>^!ú$� �B�Ò>g3>¡OD$¯°;/�¯~j>uv7rÅe�L��*� �
GGGG�
����°±²³�1. ��S±Ý���Mo� Web"¡y>{�»――+þeÌþ_`�Webèr=EÑ�Ø� q^ ¢µ£¤2NIKKEI NET^IT¥¦pj� 2005/4/1
����ª�«�s ÄÅ[«Æ��`a�i´]ÇkÈ°��ÉÊËÌ�ÍÎËÌ�ÏÐ�»��9ÍÎ7��ÑQRS���BÒ/¸Ó34Fs
1. §¨'�, PSB01 Hawaii, 2001
2. �AD�¡��'�, ISMB, Copenhagen, 2001
3. �AD�¡��'�, HLT, San Diego, 2002
4. �AD�¡��'�©GÀÖ7ÓÔשGÄ9Ó8:Ó, 2002G
5. §¨'��ÃÀªÂ©GRome/Frascati, 2002
6. §¨'�, LangTech, Berlin, 2002
7. 84'�, IEEE Int. Conf. of NLP and KE, Beijing, 2003
8. §¨'�, WS on Multi-layered Annotation, Bielefeld, Germany, 2004
9. 84'�, LREC, Lisbon, 2004
10. §¨'�, EBI, Cambridge, 2004
���������� ���������
���
11. 84'�, ICON2004, 2004.
12. 84'�, SMBM, Cambridge, 2005
13. §¨'�, BIT 2005, Tainan, 2005
�
�*��ª�«§´µ¶�
� <=C> ��&�Ú��|0¸îVP¤ »��î����£��o5�Q¤
� G����»®�½G��� �}~}��]�����}]z{��¤å5�Q7: G���
&�st¸H�����¸H ��°��� æ£��<=�ÆÇL6:M�N�
.G5�QG���;XP�M�������äåæ£:ÆÇL6:MÌ� ��
�� �67�
� : G���&��Õ'�����ML6 }i¡]�}h]} ©ª ?§�� ^P�
89\í<=¢Á¬îï�;�l�Qjäw�0�¯�-����¸H���m��
Á¬���M����¸H�� ü"�V6��-��3nè�i��äåæ£: �
��� ÆÇ{M�±} /���à� ^P�í<=¢Á¬îï� <=C>\�
�}~}��]�����}]z{��¤å5�Q:M&�í<= Zq\°�����}>� �6s
8M1U>�67�
�
�� �����������
�D��·<����={��¤�7^¯�
ÔÕÖs \×s Øus _ÙÚÛ ÜÝs
���º�G
���«qG
���qG
�Ö;¬�6Ö8G ÖÔG =9>;97G
@9Ô×>9×:G 5;Ö®:��ÓÔ×G
9Ô¯G°ÔÖ7Ö×?G�>Ó7¯ÓÔ×G
±Ö;G�ÓÖ7Ö×?G
���D¬
�ÞQ$G
G
ÚCpG
<Ia)>Sîã²�Sî7'®ë
ؼ�>�æ$�#��«³ ��
���´µ¶@!ú$�¬D®D$
7Üõ�Îe'E�·>57��
Y�ÎgSî>!ú$�7�óN�
�klhid6yÅeK¶std
¸·;�¹º�ÎeþØ»G
^WD#EF�Xã²���;K¶
kl�\7K¶st>g3>!ú$
�®D$>uv>Sî7'ÔØ�
�>uv«¼d½¾ëØ;;�7�
��V7t¿Y>6À�UëØ
K¶st>g3>!ú$�¬D
®D$>uv7�ÁõÅeþ:g
37�Sî>��Â;>uvTc'
Dd�÷;dxV;ëØ»G
���º�G
����qG
ÄÅ2E�ß�à£Æ
óN��kl;ÇÈ£
����DG
���D¬
�ÞQ$G
G
��� pG
@{>WD#EF�X7É#>Å
2E�ß�ࣻ<Ia)Sîdr
Ê;ÅgóN��klOÆ���
�D¯°7'Åe�O�Ë>�²V
ÌuvÌd�Ì;Åe��P)�
a)uvq"òuv�ÌÍ;G#
ã2#»G
���������� ����������� ��
���
����G
����qG
G ��«qG
=ÄÄ ?@ABCDEF
�a)8!uvq;>
WD#EF�XG
��DQ
ÎçG
G
¹� pG
ËÌ)Í7'ÅeªÏVÌuvd
ÐeþØ =ÄÄ ;���>¯~kl
d�Ñ7uvdÐeþØ}~ Ò
;�uv�cdÓ�ëØ·;ã�û
Ô>uvdÕUëØ·;dxV;
ëØG
����G
¹���qG
�:?ÖÔ¯G �6977ÖÖG
9Ô97?�:�×ØÖ;Ù97Ó�Ù�G
9Ô¯G �9Ó�Ó®97G
ÙÖ¯:7ÓÔ×G ±Ö;G ¯::8G
9Ô97?�:�G
�OG
Ù$ÚGº� pG
���ºë4¨�d�*>¯~VÌ
ÛÜ;�ZVÌÛÜ;7�Ô�*Î
ÝÎd)VÌÞßãànØ·;�
ág�*>Ld>ÛÜd�XÅgÏ
¬�d¯°ëØ·;²�âã>��
7'ëØ)Vl¼d�nØ@ã
ÙÈÌuv;ÌÐeþØ»̂ WD#
EF�Xã²�äQ����>�²
VÌuvÒ�DX'×Jå�>Ò
�DXèruvü²d¸·;ã�
·>Sî>е_`dÓ�ëØ·
;xV;ÅeþØ»G
����G
����«qG
���qG
=ÄÄ ?@ABCDEF
�a)8!uvq;>
WD#EF�XG
æpç��
è�é!�
G
�º pG
ËÌ)Í7'ÅeªÏVÌuvd
ÐeþØ =ÄÄ ;���>¯~kl
d�Ñ7uvdÐeþØ}~ Ò
;�uv�cdÓ�ëØ·;ã�û
Ô>uvdÕUëØ·;dxV;
ëØG
�����G
¹���qG
:×4�Æ��D�BE
�!"æ2µÅþ<�
)>2üG
��P)G
��éD��Ú� pG
4�Æ���� !"#$>É#
WD#EF�X;Åe��P�a)
uvq"òuv�ÌÍ;G#ã2
#»G
�
�����¸¹��º¯�
Þs \ uß�àá(s Íâ�ã±s ä�ås ä�æç
ê9;¬GÒÖ6Ô�ÖÔG
¤QOæà�P)G
��¥G
{nVÌ��kl� ;�ZVÌ
ËÌ)ÍϬ�;>�X7'ëØ
G3uv7'ëØëìX��d
¸G
��P)P)�G
_`lJ)¿uvaG
}~uv�G
����G
¹�GÛq�
º�GºqG
í;9îÓÔ¯GØïGÒÖ�6ÓG
¤QO}�E�4B�
P)G��¥G
ÄíÒ>g3>¯�ýþ� �'×J
<Ia)Sî7'ÔØ�õ_`k
l7'ëØuvëìX��d¸G
��P)P)�G
_`lJ)¿uvaG
}~uv�G
����G
¹��Úq�
¹��ÛqG
ÃÖ±Ó9GíÔ9ÔÓ9¯Ö>G
¤ OX!ú$� �
B�Òq�DGGGGGG
ðqÃ¥G
^SÝ<�)>K¶>¯~Yj7'
ëØG3uv>üñ>g37�}~
ÒãuvëX�d¸»ágOwE
�ß�à£ò:×4�Æ��D�BE
�!"æ2µÅþ<�)>2üó7
euv�cdü²Åe�r¸G
��P)P)�G
_`lJ)¿uvaG
}~uv�G
�����G
¹�G«q�
¹��ºqG
���������� ���������
���
ô¯ÖGõ96ÔG
G�u�vG��yP)�
��G
^äåóN��klj7'ëØG3
uvëX�>g3»ágOwE�ß
�à£ò:×4�Æ��D�BE�!
"æ2µÅþ<�)>2üó7eu
v�cdü²Åe�r¸G
��P)P)�G
_`lJ)¿uvaG
}~uv�G
�����G
¹�G«q�
¹���qG
�
�
�� ���
�3456789:;�<�
í<= ��&�st¸H�ËÀ¸H�£�¬� v2�Qî����±�¬î����
-. Zq!"L6:M;�á�-������ «²M¸H�¼�Ä: �� 3
ß��L6�����GL6:M;X�Q7!Q�|0Ö <=�ÖHI�ÏE�
� <=�#Iâo0ÖM�Zq !"Mì"Ñ�CDo��â Ú0Ö�±º�
�u�u X6<=�#:M�l�Q7�
: ��M��&�í¢Á¬îï�\�HI�ÏE���!"���[ �r��^
� W��u�)L6ZqXÉQ:MO>���-Zü;X�QM1U �67�
¢Á¬îï���¡�Ö �dW�v2�Qî�� £�¬� ���st¸H� å
æ¢Á½[�m��Á¬�� äåæ£���ÊÑ�Q,-�m��Á¬�M�#Z
q!"��Ðy²tM�#��-����¸H���®Âº �#!"~��¿~
L6-.��±} ���;X�Q7�
�34�2=>?�
� í¢Á¬îï�&�HIN-E�;&�\[^Ãòóôõ�}6�ö�\'^EF+G
Í EF�Ç�Q�Ô÷��woxÐz{�\Y^�Ô÷��st¸Håæ��çuè
[:Ý%ÒÛ?eNêÝ%\oxyz^�M�#ÃÄ�Å °�cÇWpêÝ%ef
�Q7:�> HI&��� �d-<=½��¢\���[++ Q}��ÇL6-
.��� I ;ÊÇ5� �67�
� !Q�ÏE���<=µ¶ efM� &�\[^§�¡£ ÓØ�ÓOF- åæ���
[\�hÈ�^ ef\oxÐz^�\'^?eN-©�l§¨;IYL6ÓØïÁ��wox
|z{�\Y^�¦©�l§¨ ���Ç�# £�¬� v2�Qî� £�}wox|z{�\�^
ÓOF-��Çt�À���[ws_{\oxQz^�\�^ABCDMEF+G Q}
hÊ+GM� ghi ¢��x¡ãwoxÚz{�\�^äåæ£ m��Á¬�M�úp�
_�Þ� YZv£��¬ã m��Á¬�M_�Þ�wwoxyz{M�#ZqXÉQ7:
�>&��r�W��� ÏE��;�� ûÓT_ W ;X67�
QMU&��hÈ� Fm&�¸HØÙ;&*�� ��ûÓ;XP�MÙ D;W
±�«�¡�£�0�¯u�¬�£��ã� �H���¦¼��£M�#�: ��;
®JN-½��¢ W M�ò�X6�&���4ð W M-� �67!Q�£��¬ã
_�Þ�&�67MË L�;äåæ£ Ü÷p�_�Þ� ûÓ WM%-5���
�;Úy�� ½��¢�V� �±� �67�
���������� ����������� ��
���
í¢Á¬îï� Zü&�ðñ V#-HIMÏE��� �N-Zq\7>�Q:
M¹º;&-�7�ú�CÏ-:M&�:�> Zq\�lN�@�5� �wox¸
z{ !"~��woxQz{ !"è�i����[-.�ÏE���HI��L6
Ké<=¹º;&üZ;�-�ZqXÉQ:M;X67�
:�> Zq&�¢Á¬îï� �â��M���ûÌ0Ù�7>�QW ;XP�
� ]µO> ��&!¹�!� �-�7���!"~�;7>�Q�67-Ü÷
�����!"è�i����[\��� <=>²³�~¼�Ä äåæ£>�.
V#��±� ��O\�¢Á¬îï���M� &CÏM-67�
¢Á¬îï�u�"�M� &�� �)Ç�ï-��efM&�-P�HI�
ÏE��O>!"���[:M�#�¡�~z}¤�z�];�Op�£}h}��]-����\�G
;�Q:M���� �67!Q�ø�����¸H��ÏTM&PG�ûI�~�
��:�WM���ÏT�²B�QËÀ¸H����c�#M�#¢Á¬îï�
1Uj\�~Çðëm;X6:M~s;�Q:M&��� �����ËÀ�ý�
�Q�����[ <=���-´µTU67�
�@A�34�BC�
� �ý ¢Á¬îï�;&�+GN�<Ú-st¸HwÛtãO>Ü÷+G©�L6
åæ ¸H�-.{ ��\Ï��~ÇN-Fmà� ~�;�6:M�!Q��
¿qm��Á¬�����ËÀ¸HM¿×�º6�í ��ef�Q7Q¹�<
= ��&�ÛtãO>Ü÷+G©�L6¾´�CD\^O��Ü÷+GMËÀ¸H
M¿×pº �VPÓÙ-�I¸H�#�&>-O�Q7: µ� <=�#
Q}�&��
\[^����O> �67m��Á¬� wâ{�I++�
\'^ ¨�4UQ D¸H�
\Y^¥½ �67µ¶¥�L6:Mëm�L6�¦©�l§¨�
\�^ÃñóN-�I��MñóN�I��!"L6���
�p� 5>-6<=\ÎÏM-67�ý ��¡�<= Zq&�:�> ��»
��pO�ýD Qf_<=�Ê�¡\�67�
!Q�¾.|0Ö Ö��¡}¤�hz�]�ª}��}i¡]�}h]}���z��£����äåæ£�^º6
�}Þz�¢�h�h�M�#�í¢Á¬îï���±\ø�Ê8��\��5��<=\ôØ�
�67í¢Á¬îï� Zq&�����¸H øDO>�:�> ����PL6
:M�-67�
�DEFGHI�JK�
�� Ké<=M¢Á¬îï� ���±�
¢Á¬îï� ë�;&�KK <=> P�F�HIN-�k5\ÏT5�6H
I�ÏE��N-<=M��67-=�M²³<=; %ëmM-6!"N-<=M
¡���W�MW�Ü��Q7�
²³;�#���[ef¹º�N�L6M�g�O�p�����Q�FÇF -
����[ef�̱P��6F�P�F�HIN½"FT}6<=>²³ ¢Á¬
���������� ���������
���
îï�M� &ê��-W M-67F��KK<=�CDo�L)6M��� �
@F\���D� -�KéN-<= ²M-P��67-=��#¢Á¬îï�
M� Ü»\-�-67�
: D;�HI�ÏE����67-µ¶++�CDÌ 6I�M�!"Ñ�l
L��M�#�u�upºQë�&�ÃÄ��O�QMO#7�
Y� <=»� Ì<=> ¡Z�
: � <=;&�þ¢-K� �rM�<=£� Q} µ¶++\CÏM1U�
:�>: =�;�ûþ®M�Q7¤ �)®�O> Ì<=>p}�Ä��x
¸- Ä¥<=>¦Ç�Q7!Q�£��¬ã_�Þ� ++M�¦©�l§¨ ½úM
�#uv�� ++�W�KO»M�ú�=���Q7: ¿q�HI�ÏE���
uv�� ��OPM�<�Q§¨;����M!"���[ ++\;�Q7�
!Q�¢Á¬îï�\3n�QKN�µ¶N-§¨&� Ì<=>�M� W§¨N
-<=§¨;XP�d>&�<� þ�QZqðÉ���B� ��;�� <=
u�)L6<=>M-� �67!Q�¢Á¬îï�¦Ç ©°£äW�þ¢-<=>
\kÙÓ��6M�#§¨;��� þ�QZqXÉQ7�
ª� ¢Á¬îï�Ý� ¹��
¢Á¬îï� ¾´;�m��Á¬�£ �ð�«¬\)®� «¬�!Q�±�¬
î��<=� ïð«¬\Ï�£:M��B�® 5�����û#Ý�¹�M<=
�� ¹���Q7�
m��Á¬�£ u�"�&���Çt�À-.�m��Á¬�<= ���;X6
�ó«¬wÎÏ�£{�^¯��Q7«¬ °÷&�����O> m��Á¬�++
;XP�¢Á¬îï�M ½"FWV���C&¢Û-��±�Q7�
!Q�IâÌÍÕD; n��;�ï𫬠ðÃñóMñó ��ñ�ð�FM®
���2î��ñ <=��&�P�F X6��±} Ó�Ké��;XP�Ï�®
�WÊ�f�^¯�L6:MM�Q7«¬�&��� !"N-<=;&�¢Á¬î
ï�±;"±é²;�N-�s�Q¹��M.}�KéÏE<=Ê�f�^¯�
L6:MM�Q7: b`&�±�¬î�� £ Ké<=Zq Ë Ó5M�� <=
\�� ����¸H; § ¬�TU6´µ ��51U6M�M��W ;X�Q7�
�34�2�LM=NO�
� í¢Á¬îï�;&�] <=>²³MWjk-@³W;�Zq]µ�°}6:
M�´Ü�Q7M���äåæ£>M&�¥�<=� �(�ï���¢ ¥�eµwÚ
ý{-.�jk�@³M�Q7¿qM� �äåæ£ Q} ����»®¼�½M
�#�������� <=> �ý²}6:M�Zü�Q7!Q�Ú��y0Ù�
eµ�Qðø� åæñ��L6(�ï���¢W�í¢Á¬îï� �m ?p;X
6ð��;PM�Qst£HI����st¸Hñ�ýDX �:�!;P��<=
�� �QH���¦¼��£�Ö¶·¸�±�«�¡�£-. <=½��¢
¥½ <=��à�Q²³�L6:M��P�Q7�
���������� ����������� ��
���